Upload
others
View
5
Download
0
Embed Size (px)
Citation preview
Florine Duval
New applications ofthe interaction betweendiols and boronic acids
New applications of
the interaction between
diols and boronic acids
Florine Duval
Thesis committee
Promotor
Prof. Dr H. Zuilhof
Professor of Organic Chemistry
Wageningen University
Co-promotor
Dr T.A. van Beek
Assistant professor, Laboratory of Organic Chemistry
Wageningen University
Other members
Prof. Dr W.J.H. van Berkel, Wageningen University
Dr J.A. Peters, Delft University of Technology
Dr A. van Amerongen, Food & Biobased Research, Wageningen UR
Prof. Dr A.J. Minnaard, University of Groningen
This research was conducted under the auspices of the Graduate School VLAG (Advanced
studies in Food Technology, Agrobiotechnology, Nutrition and Health Sciences).
New applications of
the interaction between
diols and boronic acids
Florine Duval
Thesis
submitted in fulfilment of the requirements for the degree of doctor
at Wageningen University
by the authority of the Rector Magnificus
Prof. Dr A.P.J. Mol,
in the presence of the
Thesis Committee appointed by the Academic Board
to be defended in public
on Friday 9 October 2015
at 4 p.m. in the Aula.
Florine Duval
New applications of the interaction between diols and boronic acids,
140 pages.
PhD thesis, Wageningen University, Wageningen, NL (2015)
With references, with summary in English
ISBN 978-94-6257-471-7
To the memory of my father,
builder with fascination for science
Table of contents
List of abbreviations
Chapter 1. Introduction .................................................................................................... 1
Chapter 2. Key steps towards the oriented immobilization of antibodies using boronic acids ............................................................................................................................... 15
Chapter 3. Design and synthesis of linkers for the oriented and irreversible immobilization of antibodies ................................................................................................................... 29
Chapter 4. Sensitive thin-layer chromatography detection of boronic acids using alizarin ........................................................................................................................... 59
Chapter 5. Selective on-line detection of boronic acids and derivatives in high-performance liquid chromatography eluates by post-column reaction with alizarin ........... 71
Chapter 6. General discussion ........................................................................................ 93
Appendix. Information supplementing chapter 5 ........................................................... 105
Summary .................................................................................................................... 125
Acknowledgements ..................................................................................................... 127
About the author’s chemical pathway ........................................................................ 129
Publications ................................................................................................................. 130
Overview of completed training activities .................................................................. 131
List of abbreviations
2D-TLC two-dimensional thin-layer chromatography HOBt hydroxybenzotriazole
AACPBA (4-allylaminocarbonyl)-phenylboronic acid HPLC high-performance liquid chromatography
Ab antibody HSA human serum albumin
Abs antibodies IgG immunoglobulin G
ABTS 2,2'-azino-bis(3-ethylbenzothiazoline-6-sulphonic acid) LOD limit of detection
ACN acetonitrile mAU milli arbitrary units
APBA MeOH methanol
ARS mg milligram(s)
Boc MIDA N-methyliminodiacetic acid
Boc2O
3-aminophenylboronic acid
alizarin Red S
tert-butyloxycarbonyl
di-tert-butyl dicarbonate min minute(s)
Boc-Cys N-a-t-butyloxycarbonyl-L-cysteine mL milliliter(s)
CHES 2-(cyclohexylamino)ethanesulfonic acid mmol millimole(s)
CPBA 4-carboxyphenylboronic acid MPBA 4-methoxyphenylboronic acid
DCM dichloromethane NCBI National Center for Biotechnology Information
DEAM-PS diethanolamine-functionalized polystyrene resin NHS N-hydroxysuccinimide
DIPEA N,N-diisopropylethylamine NMR nuclear magnetic resonance
DME 1,2-dimethoxyethane PBA phenylboronic acid
DMF dimethylformamide PBS phosphate-buffered saline
DMPA 2,2-dimethoxy-2-phenylacetophenone PEEK polyetheretherketone
DMSO dimethyl sulfoxide PS polystyrene
DPBA diphenylborinic acid 2-aminoethyl ester PTFE polytetrafluoroethylene
DPPH 2,2-diphenyl-1-picrylhydrazyl RM reaction mixture
EDC.HCl N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide HCl RP reversed-phase
ESI-HRMS electrospray ionization high-resolution mass spectrometry RT room temperature
Et2O diethyl ether S/N ratio
EtOAc ethyl acetate tBu
EtOH ethanol TEA
Fmoc fluorenylmethyloxycarbonyl TFA
g gram(s) THF
h hour(s) TLC
HEPES 4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid UV
signal to noise ratio
tert-butyl
triethylamine
trifluoroacetic acid
tetrahydrofuran
thin-layer chromatography
ultraviolet
1
Chapter 1 Introduction
Chapter 1
2
1 When boronic acids and diols meet
1.1 Theory
Imagine a dancefloor full of men and
women, moving alone in all directions.
When a man and a woman bump into
each other, if the chemistry is right, then
chances are that they start holding hands,
making them tightly bound to each other.
Chances that a man and a woman
form a dancing pair depend on many
factors, which are among others:
Ambient conditions
Chances that a man and a woman
bump into each other are higher when the
dancefloor is busy and people are
moving a lot, than when the dancefloor
is quiet and people are standing still.
Affinity If the man’s arms are too far from
each other, or pointing in different
directions, then the woman might not
be able to hold both his hands.
Imagine a solution of diols and boronic
acids, moving alone in all directions.
When a diol and a boronic acid bump into
each other, if the chemistry is right, then
chances are that they react together,
making them tightly bound to each other.
Chances that a diol and a boronic acid
form a boronic ester depend on many
factors, which are among others:
Ambient conditions Chances that a diol and a boronic acid
bump into each other are higher when the
solution is concentrated and molecules are
moving a lot, than when the solution
is diluted and molecules are standing still.
Affinity If the diol’s OH groups are too far from
each other, or pointing in different
directions, then the boronic acid might not
be able to react with it.
Introduction
3
Interactions between men and women are far more complex than interactions between
boronic acids and diols, and outside the scope of this thesis. From now on, therefore, only
boronic acids and diols will be discussed. A boronic acid B0 has a vacant valence orbital,
which makes it a Lewis acid. By accepting a
hydroxyl group, the conjugate base B– is
formed (Figure 1). In presence of a 1,2-diol
L, studies have shown that mainly, B0 first
reacts with L to form the corresponding
boronate ester B0L, which is then converted
into the more stable B–L (in aqueous
solutions).1 These reactions are reversible
and, as differently explained above, the
equilibria depend on many factors. In the
next page, more of these factors, specific to
boronic acids and diols, are explained.
To hold a woman’s hands, the man needs
to point both his arms towards her.
The woman needs to be attractive enough,
so that the man wants to hold her hands.
Competition If there are many women around the
dancing pair, or more attractive ones,
chances are high that the man goes with
another woman.
If there are many men around the dancing
pair, or men with more inviting arms,
chances are high that the woman goes
with another man.
To react with a boronic acid, the diol needs
to point both its OH groups towards it.
The boronic acid needs a low pKa,
so that the diol wants to react with it.
Competition If there are many boronic acids around the
boronic ester, or more acidic ones,
chances are high that the diol goes with
another boronic acid.
If there are many diols around the boronic
ester, or diols with a better conformation,
chances are high that the boronic acid goes
with another diol.
Figure 1. Possible equilibria in aqueous solutions of
boronic acids and diols.1
BOHHO
BOO
B- OHHOOH
B-OO
OH
+–H2OH+
+–H2OH+
– H2OOHHO
+
––– H2O
OHHO+
B0 B0L
B–LB–
L
L
Chapter 1
4
pH, diol pKa and boronic acid pKa. The optimal pH for boronate ester formation, in
general, is about half-way between the pKa of the boronic acid (e.g., 8.70 for phenylboronic
acid) and the pKa of the diol (e.g., 12.38 for glucose).1 To enable binding at neutral pH,
therefore, efforts have been made in lowering the pKa of boronic acids.1
Boronic acid structure. The structure of the boronic acid influences its pKa significantly,
and hence also its ability to bind with diols in given conditions.1 For example, if
phenylboronic acid is considered, its pKa is made higher by electron-donating groups (e.g.,
methoxy-) and lower by electron-withdrawing groups (e.g., trifluoromethyl-) on the phenyl
ring, affecting the stability of the resulting boronate ester. Moreover, bulky groups close to
the boronic acid may hinder the binding with diols.
Diol structure. A more favorable geometry of the 1,2-diol normally results in a higher
affinity for the boronic acid.1 For example, catechol has a perfectly planar cis-1,2-diol,
whereas sorbitol, although it contains several 1,2-diols, needs to rotate away from its ideal
conformation in order to create a planar 1,2-cis-diol. Sorbitol, in turn, is able to create a
planar 1,2-diol, whereas ribose cannot freely rotate the relevant C-C bond and can thus only
display a near-planar cis-diol. These differences directly result in differences of association
constants with phenylboronic acid (Figure 2).2
Figure 2. Association constants (Keq) with phenylboronic acid at pH 7.4, 0.10 M phosphate buffer2
Competition. The proportions of boronate esters that are formed strongly depend on the
composition of the solution. For example, in a mixture containing phenylboronic acid and
equal amounts of catechol and D-sorbitol, more phenylboronate esters of catechol than of
D-sorbitol will be found. However, if a large excess of D-sorbitol is added, more
phenylboronate esters of D-sorbitol than of catechol will be found.
OH
OHHO
OH
OH
OH
OH
OHO
HO
OH
OH
HO
830 370 24Keq (M-1)Catechol D-sorbitol D-ribose
Introduction
5
1.2 Well-explored applications
The boronic acid-diol interaction has found many applications. A few of them are
highlighted here, some of which are illustrated in Figure 3 (next double page).
Many boronic acid-based sensors for sugars have been developed, with efforts in making
these sensors selective for specific sugars. For example, particular attention has been given
to the selective and sensitive detection of glucose for the monitoring of glucose levels in
blood of diabetic patients, and the development of glucose-responsive materials that
automatically release insulin when glucose levels are too low.3-5
Glycans, formed by several monosaccharides attached to each other, are found in many
proteins (thus glycoproteins). Various materials that contain boronic acid moieties have
been used for the reversible capture (≈ fishing) of glycoproteins from complex mixtures,
facilitating their detection, detailed analysis or purification. Glycoproteins have also been
immobilized on boronic acid-functionalized surfaces, enabling further applications that rely
on the properties of these glycoproteins (e.g., detection of antigens through the use of
surface-bound antibodies, see section 2 of this chapter). Furthermore, as glycans are also
found on cell surfaces, boronic acid-containing materials are being developed for the
detection or capture of cancer cells, which overexpress certain glycans at their surface.6-9
The other way around, the boronic acid-diol interaction has been exploited for the detection
of boronic acids at the surface of newly developed materials (e.g., for the applications
above).10-12 Alizarin red S (ARS) is widely used for this, as it shows low fluorescence in its
free form, and high fluorescence when it is bound to a boronic acid via its catechol moiety
(see chapters 4 and 5).
In organic synthesis, diols are used for the protection of boronic acids,13 and boronic acids
are used for the protection and activation of diols, and for the catalysis of reactions
involving the formation of diols.14 Furthermore, diol-functionalized materials have been
developed to function as protecting groups of boronic acids and facilitate their modification
and isolation (see chapter 3), and vice versa.14
Chapter 1
6
Figure 3. Some known applications of the boronic acid-diol interaction.
Healthy cell
Healthy cell
Cancer cell
Diagnosis of cancer
Glucose monitoring for diabetic patients
Capture of glycoproteins for their analysis
Blood fromGlucose
Boronic acid-functionalizedmagnetic nanoparticle Fishing out
with a magnet
Non-glycoprotein
Glycoprotein
diabetic patient
Diol
Boronic acid
Analysis
Introduction
7
Figure 3. (Continued).
Protection of boronic acids
Cleavage
Diol-functionalizedpolystyrene bead or
Diol-containingsoluble molecule
Protection of diols
Detection of boronic acids(see chapters 4 and 5)
Other chemicalreactions
Boronic acid-containing molecule
Boronic acid-functionalizedpolystyrene bead or
Boronic acid-containingsoluble molecule
Cleavage
Other chemicalreactions
Diol-containing molecule
NewBoronic acid-
containing molecule
NewDiol-
containing molecule
Immobilization of antibodies for sensing(see chapters 2 and 3)
Antibody
Boronic acid-functionalized material
Alizarin or alizarin red S
Orangefluorescence
(Almost) no fluorescence
Boronic acid-functionalized material
Chapter 1
8
2 From diagnosis of depression to boronic acids
If you feel depressed, you may go to your doctor, who will ask you many questions. Your
answers will indicate whether you actually suffer from depression, and which type of
depression you may have. This way to diagnose depression is subjective and may lead to
inappropriate treatments. This lengthens the disease, or might even fully hamper a cure.
Knowing this fact, Brainlabs B.V. started a project for the development of a device that
would enable the diagnosis of depression through urine analysis. Instead of asking
questions, the doctor would put a urine drop on a chip, wait for a few minutes and read the
results. This multidisciplinary project involved several companies and institutions, who
worked on different aspects of the project: finding which biological markers (= biomarkers)
in urine would give the best evidence of the different types of depression, performing
clinical studies with depressed and non-depressed people, making suitable chips…
Functionalizing these chips to enable sensitive detection of the biomarkers was the mission
of Wageningen University, in particular my PhD project within the Laboratory of Organic
Chemistry.
How to functionalize a chip in such a way that it is able to detect a biomarker? One way of
doing this would be by immobilizing specific antibodies in a specific way on the surface of
such a chip. Antibodies are, roughly seen, Y-shaped proteins that bear two binding sites
(Figure 4, top left), regions that recognize and bind to substances such as bacteria, viruses,
toxins, hormones or other molecules (= antigens) in a specific way: each antibody is
specialized in one antigen. Thus, one popular way of detecting a particular biomarker (=
antigen) for diagnostic purposes is by performing an immunoassay. The use of such an
assay can, for certain classes of assays, be described as follows (Figure 4, top right): 1)
immobilizing the corresponding antibody on a chip, 2) incubating this chip with the sample
(e.g., urine), which results in binding of the biomarker (= antigen) to the antibody, 3)
incubating the chip with a solution that contains a fluorescently labelled antibody, which
binds to the biomarker, 4) measuring the fluorescence on the surface of the chip, which is
related to the concentration of biomarker in the sample.
Introduction
9
Figure 4. Antibody structure, immobilization (not exhaustive), orientation and immunoassay principle.
To diagnose depression, biomarkers would need to be detected in a similar manner at very
low concentrations (~ 1 pM) in urine. One key factor of the method sensitivity is the way in
which antibodies are immobilized on the surface of the chip: statistically, if very few
antigen molecules are present in solution, many antibody binding sites need to be available
for capturing at least some antigen molecules. The number of available antibody binding
sites depends on the number of antibodies immobilized on the surface (related to the
efficiency of the immobilization procedure), their integrity (which may be affected by harsh
reaction conditions), and their orientation.
Various strategies have been developed for protein (thus antibody) immobilization.15,16
Physical adsorption is the most simple and widely used way (Figure 4, bottom left): when a
hydrophobic support, such as a polystyrene plate, is incubated with an antibody solution,
the antibodies spontaneously stick to the surface via several of their own hydrophobic sites.
However, the antibodies may have difficulties keeping their original shape, and they may
be released from the surface again. These problems, resulting in loss of available binding
sites, can be solved by covalent immobilization strategies (Figure 4, bottom). The most
Via boronic acids
Fc region
Antigenbinding sites
N-glycans
Antigen
Antibodyconjugated withfluorescent dye
Antibody immobilization
Immunoassay
By physical adsorptionBy random covalent coupling Via antibody-binding protein
Antibody structure
Chapter 1
10
common one relies on the reaction between an activated ester on the surface (e.g., N-
hydroxysuccinimide ester) and an amine on the outside of the antibody (more precisely one
of the side chains of lysine residues), to form a stable amide bond. The antibody is attached
to the surface in a stable way, and if the conditions are right, it also conserves its shape.
However, both described immobilization strategies have a serious limitation, namely they
are “random”: there is no control over the orientation of the antibody, which may stand
upright, stand upside down, lie flat or lie on its side. Thus, statistically, many binding sites
will not be available for antigen binding, hence the sensitivity for antigen detection will not
be optimal. To solve this problem, strategies for “oriented” (or “site-specific”)
immobilization of antibodies have been developed. For example, an antibody-binding
protein (protein A, G, A/G or L) is immobilized onto the surface. When the surface is
incubated with the antibody solution, the antibody-binding protein captures the antibody via
its Fc chain only. As a result, both binding sites are available for antigen binding (Figure 4,
bottom). Although antibody orientation is optimal with this strategy, the expensive
antibody-binding proteins are immobilized on the surface in a random way, resulting in
relatively few sites that are available for capturing the antibodies. Other oriented antibody
immobilization strategies, however, do not depend on additional proteins, and take
advantage of the sugars (N-glycans) that are present in the antibody Fc chains. It is possible
to oxidize the diols in the N-glycans with periodate, and let the obtained aldehydes react
with amines on the surface, resulting in oriented and covalent antibody immobilization.
This method, however, requires prior treatment of the antibody with periodate, which might
result in degradation of the antibody. Alternatively, the N-glycans can be modified via a
series of enzymatic and chemical reactions, to provide an antibody derivative that can be
clicked onto a surface in a well-defined manner.17 This method yields highly defined
antibody orientations, but does require extensive antibody modifications. Fortunately, there
is a simpler (although less strongly binding) alternative to immobilize antibodies via their
N-glycans (thus via their Fc chains), namely to let these N-glycans react with surface-
bound boronic acids (Figure 4, bottom right).
Introduction
11
Antibody immobilization through the reaction between boronic acids on the surface and
diols in the antibody N-glycans would present several advantages for the detection of
biomarkers for the diagnosis of depression: 1) it is cheaper to modify surfaces with boronic
acids than with antibody-binding proteins, which is important for the manufacturing of
disposable chips, 2) no prior modification of the antibodies is required, making it easy to
immobilize many different antibodies on the same surface, 3) and most importantly, the
antibodies are immobilized via their Fc chain, leaving their binding sites available for
antigen binding and resulting in a high sensitivity for the detection of biomarkers at low
concentrations.
3 From one chapter to another: outline of this thesis
While biomarkers for depression were being investigated by the project partners, our work
focused on the development of a general strategy to immobilize antibodies on surfaces via
boronic acids. From preliminary experiments and gradual “discoveries” from the literature,
we became aware of the many challenges that this immobilization strategy faces, giving
rise to chapter 2, which discusses several important points that need to be taken into
account when one plans to immobilize antibodies via boronic acids.
One big issue being the reversibility of the reaction between boronic acids and diols, hence
the possible release of the antibody from the surface, we designed and synthesized boronic
acid-containing linkers that would enable the oriented and irreversible immobilization of
antibodies. This work is described in chapter 3.
While synthesizing boronic acid-containing linkers, analysis of the reaction mixtures was
inconclusive, as it was difficult to see which spots on thin-layer chromatography (TLC)
plates corresponded to boronic acids. To solve this problem, we developed a new TLC
staining method based on the reaction between boronic acids and alizarin, resulting in
fluorescent spots where boronic acids are present. Chapter 4 presents this work in detail.
Chapter 1
12
Although TLC is very useful to synthetic chemists, analysis of reaction mixtures by high-
performance liquid chromatography (HPLC) is sometimes necessary for more accurate
information or for optimization of preparative HPLC conditions. Chapter 5 presents the
development and applicability of a method for the on-line HPLC detection of boronic acids
using alizarin.
Looking back, some useful findings have been made, some things may have been done
differently, and some results may be the starting point of further developments. Chapter 6
provides a critical discussion on the work described in this thesis.
4 References
1. J. A. Peters, Interactions between boric acid derivatives and saccharides in aqueous media: Structures
and stabilities of resulting esters, Coordination Chemistry Reviews, 2014, 268, 1-22.
2. G. Springsteen and B. Wang, A detailed examination of boronic acid-diol complexation, Tetrahedron,
2002, 58, 5291-5300.
3. S. Jin, Y. Cheng, S. Reid, M. Li and B. Wang, Carbohydrate recognition by boronolectins, small
molecules, and lectins, Medicinal Research Reviews, 2010, 30, 171-257.
4. J. S. Fossey, F. D'Hooge, J. M. H. van den Elsen, M. P. Pereira Morais, S. I. Pascu, S. D. Bull, F.
Marken, A. T. A. Jenkins, Y. B. Jiang and T. D. James, The development of boronic acids as sensors
and separation tools, Chemical Record, 2012, 12, 464-478.
5. J. S. Hansen, J. B. Christensen, J. F. Petersen, T. Hoeg-Jensen and J. C. Norrild, Arylboronic acids: A
diabetic eye on glucose sensing, Sensors and Actuators B: Chemical, 2012, 161, 45-79.
6. H. Li and Z. Liu, Recent advances in monolithic column-based boronate-affinity chromatography,
TrAC Trends in Analytical Chemistry, 2012, 37, 148-161.
7. S. Ongay, A. Boichenko, N. Govorukhina and R. Bischoff, Glycopeptide enrichment and separation
for protein glycosylation analysis, Journal of Separation Science, 2012, 35, 2341-2372.
8. X. Wang, N. Xia and L. Liu, Boronic acid-based approach for separation and immobilization of
glycoproteins and its application in sensing, International Journal of Molecular Sciences, 2013, 14,
20890-20912.
9. C.-C. Chen, W.-C. Su, B.-Y. Huang, Y.-J. Chen, H.-C. Tai and R. P. Obena, Interaction modes and
approaches to glycopeptide and glycoprotein enrichment, Analyst, 2014, 139, 688-704.
10. A. E. Ivanov, A. Kumar, S. Nilsang, M. R. Aguilar, L. I. Mikhalovska, I. N. Savina, L. Nilsson, I. G.
Scheblykin, M. V. Kuzimenkova and I. Y. Galaev, Evaluation of boronate-containing polymer brushes
Introduction
13
and gels as substrates for carbohydrate-mediated adhesion and cultivation of animal cells, Colloids
and Surfaces B: Biointerfaces, 2010, 75, 510-519.
11. D. Zhang, K. L. Thompson, R. Pelton and S. P. Armes, Controlling deposition and release of polyol-
stabilized latex on boronic acid-derivatized cellulose, Langmuir, 2010, 26, 17237-17241.
12. P.-C. Chen, L.-S. Wan, B.-B. Ke and Z.-K. Xu, Honeycomb-patterned film segregated with
phenylboronic acid for glucose sensing, Langmuir, 2011, 27, 12597-12605.
13. C. Malan, C. Morin and G. Preckher, Two reducible protecting groups for boronic acids, Tetrahedron
Letters, 1996, 37, 6705-6708.
14. P. J. Duggan and E. M. Tyndall, Boron acids as protective agents and catalysts in synthesis, Journal of
the Chemical Society, Perkin Transactions 1, 2002, 1325-1339.
15. F. Rusmini, Z. Zhong and J. Feijen, Protein immobilization strategies for protein biochips,
Biomacromolecules, 2007, 8, 1775-1789.
16. P. Jonkheijm, D. Weinrich, H. Schroder, C. M. Niemeyer and H. Waldmann, Chemical strategies for
generating protein biochips, Angewandte Chemie-International Edition, 2008, 47, 9618-9647.
17. R. van Geel, M. A. Wijdeven, R. Heesbeen, J. M. M. Verkade, A. A. Wasiel, S. S. van Berkel and F. L.
van Delft, Chemoenzymatic conjugation of toxic payloads to the globally conserved N-glycan of native
mAbs provides homogeneous and highly efficacious antibody–drug conjugates, Bioconjugate
Chemistry, 2015, DOI: 10.1021/acs.bioconjchem.5b00224.
Chapter 1
14
15
Chapter 2 Key steps towards the oriented immobilization of antibodies using boronic acids
Oriented immobilization of antibodies using boronic acids presents a strong potential for
improving immunoassay performance but is not widely used yet, possibly because of
difficulties encountered in its implementation. How to choose the boronic acid structure and
how should it be attached to the surface? How to choose an antibody that will bind to the
boronic acid? In which conditions should the binding take place for an effective oriented
antibody immobilization? How to make sure that the antibody stays on the surface? This
chapter provides answers to these questions through analysis of the literature and personal
suggestions, and thereby intends to facilitate the development of this promising antibody
immobilization strategy.
This chapter is a slightly modified version of the following publication: F. Duval, T.A. van Beek, H. Zuilhof, Key steps towards the oriented immobilization of antibodies using boronic acids, Analyst, 2015, DOI: 10.1039/C5AN00589B.
Well-designed boronic acid surface?
Suitableantibodyglycosylation?
Properimmobilizationconditions?
Stableimmobilizedantibody?
Chapter 2
16
1 Introduction
Antibody (Ab) immobilization is an important step in the preparation of devices for the
detection of various analytes (= antigens) such as proteins, bacteria, viruses, drugs and
toxins. Antibodies (Abs) are proteins and protein immobilization strategies, which have
been reviewed in detail,1,2 can be divided into two categories: random and oriented (or site-
specific). Oriented immobilization is especially important for Abs in immunosensing
applications, as their binding sites need to be kept available for antigen binding. Using
oriented immobilization, signal enhancement factors of over 200 have been obtained,3
displaying the significance of Ab orientation. Strategies for oriented Ab immobilization,
which have recently been reviewed,4,5 aim at attaching the Ab to the surface via its Fc
region (Figure 1, right edge), far away from the binding sites. Abs are N-glycosylated (i.e.,
carbohydrate chains are attached to the amide nitrogen of Asn residues of Abs and thus
defined as N-glycans) in their Fc regions, thus a proper orientation of the antibody can be
achieved by immobilization via its N-glycans. This has mainly been done in two ways, both
taking advantage of available vicinal diols in the N-glycans: 1) by oxidation of diols into
aldehydes, which then react with nucleophiles attached to the surface, or 2) by reaction of
diols with boronic acids attached to the surface. The present chapter focuses on the second
approach.
Figure 1. Principle of oriented antibody immobilization using boronic acids.
Boronic acid- functionalized surface
Antibodyfront view
Antibodyside view
Antigen binding sites
Fc region
N-glycans
Diol in N-glycan
Key steps towards the immobilization of antibodies using boronic acids
17
Boronic acids react with 1,2-diols to form boronate esters (see chapter 1). This reaction has
been used for many applications, including the separation and immobilization of
glycoproteins (including antibodies).6 Oriented Ab immobilization using boronic acids
(through the same reaction) has been published for the first time in 2009 by Lin et al.,7 and
appeared in another ten papers since then.8-17 An overview of these articles is given in Table 1.
Paper code Main feature Ref. Ab-BA 1 First Ab immobilization via boronic acid 7 Ab-BA 2 Novel immunosensing application 8
Ab-BA 3 Novel immunosensing application 9
Ab-BA 4 Novel immunosensing application 10
Ab-BA 5 Novel immunosensing application 11
Ab-BA 6 Boronic acid on zwitterionic polymer brushes 12
Ab-BA 7 Probing of the available binding sites 13
Ab-BA 8 Direct assessment of Ab orientation 14
Ab-BA 9 Microarray of 71 Abs 15
Ab-BA 10 Novel immunosensing application 16
Ab-BA 11 Novel immunosensing application 17
Table 1. Overview of the publications that report antibody immobilization via boronic acids. The “paper codes”
are used in this chapter to quickly refer to these papers.
This immobilization method appears very attractive, as it presents several advantages over
more common ones: it leaves the binding sites available for antigen binding (oriented vs
random immobilization), it can be done in one step without former antibody modification
(vs oxidation with periodate, use of antibody fragments or enzymatic modification of the N-
glycans), and it is relatively economical (vs Fc-chain binding proteins such as protein A/G).
Moreover, several comparative studies (in Ab-BA 1, 2, 7 and 10) suggested that the highest
antigen detection sensitivity was achieved when Abs were immobilized using boronic
acids.
The number of papers reporting this immobilization strategy, however, remains limited (11
papers, four of them having been reviewed by Wang et al.6). So the question arises: where
Chapter 2
18
does its lack of popularity come from? We believe that many researchers may have tested
Ab immobilization via boronic acids, that likely problems were encountered in most cases,
and that therefore few successful studies were published. Although promising, this
immobilization strategy is not as simple as it seems. In this chapter, meant as a tutorial
review, the reader is guided through several points that need to be considered to allow
development of the full potential of this still very appealing approach.
2 Step-wise guide
2.1 Preparation of boronic acid-functionalized surface
The first step towards a successful Ab immobilization via boronic acids is a surface
modification with the right molecules.
Boronic acid structure. The interaction between a boronic acid and a diol strongly
depends on the structure of both partners.18 This interaction, hence the Ab immobilization
efficiency, may be enhanced by a boronic acid with a low pKa value, which can be obtained
by electron-withdrawing groups or intramolecular interactions (e.g., in Wulff-type boronic
acids and benzoboroxoles).18,19 To achieve a correct Ab orientation, secondary interactions
between the boronic acid and Ab should be minimized: a hydrophilic boronic acid would
help to minimize hydrophobic interactions, and ionic interactions may be prevented by
using a boronic acid that is not negatively charged at high pH (e.g., Wulff-type boronic acid
or benzoboroxole).19 3-Aminophenylboronic acid (APBA) was used in nine out of the
eleven Ab-BA papers, but this choice was never justified and, in view of the above, may
provide room for improvement.
Spacer length and structure. The N-glycans are nested between the Ab Fc chains, and
relatively far from the C-terminus/“bottom” (Figure 1, left). If the boronic acid is directly
attached to the surface without spacer or with a too short one, then the Ab may not be
immobilized or it may be forced to “lie” on the surface, which is not optimal for antigen
binding. In contrast, if a long and flexible spacer separates the boronic acid from the
Key steps towards the immobilization of antibodies using boronic acids
19
surface, then the Ab has more freedom and it can “stand”. Thus we expect that the antigen
binding capacity was relatively high when Abs were immobilized on zwitterionic polymer
brushes (Ab-BA 6, Figure 2) that had been functionalized with boronic acids.
Figure 2. Antibody immobilization via boronic acids on zwitterionic polymer brushes.12 Adapted with
permission from ref. 12. Copyright 2015 American Chemical Society.
Antifouling surface. In Ab-BA 1, 6, 8 and 10, which report the reaction of an amine-
containing boronic acid molecule with an active ester (e.g., N-hydroxysuccinimide, NHS)
or epoxide on the surface, the active esters or epoxides remaining after the reaction were
deactivated with ethanolamine. This is required to prevent Ab nucleophilic groups (such as
the -NH2 of lysine residues) from reacting with surface-bound esters or epoxides, hence
yielding a random orientation again.
Furthermore, an antifouling coating helps to minimize other types of unwanted (e.g.,
electrostatic or hydrophobic) interactions between Ab and surface. Poly(ethylene) glycol is
widely used for its antifouling properties, and zwitterionic polymer brushes prevent the
nonspecific adsorption of biomolecules even more effectively.20,21 Thus, it may be a good
idea to prepare boronic acid-containing zwitterionic polymer brushes, as was done in Ab-
BA 6 (Figure 2). Alternatively, commercially available substrates with antifouling coatings
decorated with reactive groups such as NHS or epoxide may be modified with a boronic
acid that contains an appropriate reactive group (such as APBA). One has to keep in mind,
however, that some antifouling polymers bind with boronic acid groups, possibly resulting
in a decreased availability of the boronic acid groups for subsequent Ab binding. This
Chapter 2
20
problem may be encountered with dextran, which has been used as boronic acid blocking
agent in Ab-BA 1 and 6, and with polymers of unknown structure, based on our own
experience with proprietary coatings of some commercial antifouling slides.
Characterization of the boronic acid-functionalized surface. In order to pinpoint causes
of possibly unsuccessful Ab immobilization, it is useful to know whether the surface
actually is modified with boronic acids. In Ab-BA 3, 4, 8 and 11, changes in surface
properties before and after modification with APBA suggested that a chemical change
indeed occurred, but did not prove the boronic acid integrity. In Ab-BA 2 and 6, however,
X-ray photoelectron spectroscopy (XPS) revealed the presence of boron, and in Ab-BA 6,
narrow-scan XPS data more accurately indicated the presence of boronic acids on the
surface. XPS, though, is not available to every researcher and we would like to suggest a
much simpler and more widely applicable method. Alizarin and its more water-soluble
derivative alizarin red S (ARS) strongly bind with boronic acids to form fluorescent
complexes (see chapters 4 and 5). Outside the field of Ab immobilization, boronic acid-
functionalized surfaces have been incubated with ARS, and characterization by
fluorescence imaging then allowed for an even quantitative indicator of the presence of
surface-bound boronic acid moieties.22-26
2.2 Choice of antibody with suitable glycosylation
Many different Abs were used in the Ab-BA papers. Few of them were fully described and
the rationale behind their selection was never explained. Considering the intrinsic
properties of an Ab, a key factor for its successful immobilization is the composition of the
N-glycans at its Fc chain. These N-glycans must present diols that have a suitable
conformation for binding with boronic acids. Many monosaccharides bind with boronic
acids, but when they are attached to each other, relatively weak interactions between
boronic acids and the formed glycans are enabled. Sialic acid, however, has a side chain
with a free and flexible glycol function that may bind stronger to boronic acids.18 Based on
this, the presence of sialic acids in the Ab N-glycans is highly favorable, and possibly even
required for a successful immobilization. The Ab N-glycans vary with cell line, animal
Key steps towards the immobilization of antibodies using boronic acids
21
species, culture conditions and even from batch to batch, resulting in the presence or
absence of sialic acids at the glycan chain ends.27
Analysis of Ab candidates for immobilization is therefore advisable. Ab glycosylation may
be analyzed using mass spectrometric and chromatographic techniques,28 or more simply,
Abs may be screened for sialic acids using the periodic acid-Schiff method29 (we used a
commercial glycoprotein carbohydrate estimation kit). The affinity of Abs for boronic acids
may also directly be assessed by boronate affinity chromatography using a commercially
available boronate column,30,31 although the results will also depend on the properties of the
column and the binding conditions (see 2.3).
2.3 Antibody immobilization on the boronic acid-functionalized surface
Binding conditions. Depending on the binding conditions, (glyco)proteins can bind to
boronic acids not only via the boronic acid-diol complexation, but also via secondary
interactions, such as hydrophobic, ionic19 and charge-transfer interactions.31 As these
interactions would surely result in a partly random Ab immobilization, optimization of the
binding conditions is essential. For example, the ionic strength should be low enough to
minimize hydrophobic interactions and high enough to minimize ionic interactions.19 The
pH should be low enough to preserve the integrity of the biomolecules and high enough to
minimize charge transfer interactions between boronic acids on the surface and amines in
the antibody.31 These effects are illustrated in the article of Azevedo et al. who studied the
binding of pure human IgG (antibody, glycosylated protein), human recombinant insulin
(non-glycosylated protein) and human serum albumin (HSA, poorly glycosylated protein)
by boronate affinity chromatography in various conditions (Figure 3).31 When PBS with pH
7 was used, almost all the IgG was bound, as well as a significant amount of HSA and
insulin. PBS with pH ≈ 7 was used in Ab-BA 3, 4, 5, 9, 10 and 11, in which no proof of the
Ab orientation is given. Thus, it might be that some of the Abs were actually physically
adsorbed in a random way on the boronic acid-functionalized surfaces. In contrast, Figure 3
shows a very good selectivity for IgG when HEPES at pH 8.5 was used. HEPES was used
in Ab-BA 8, although the pH was unfortunately not mentioned. This paper happens to be
the only one in which the actual Ab orientation was studied (see top of page 23).
Chapter 2
22
Figure 3. Binding of proteins to boronic acids as displayed by percentage of bound and then eluted IgG (dark),
albumin (middle), and insulin (light) in different adsorption buffers: (i) 20 mM HEPES at pH 8.5; (ii) 20 mM
HEPES, 150 mM NaCl at pH 8.5; (iii) 20 mM HEPES, 150 mM NaCl, 15 mM Tris at pH 8.5; (iv) 20 mM HEPES,
150 mM MgCl2 at pH 8.5; (v) 20 mM HEPES at pH 7; and (vi) 10 mM phosphate, 150 mM NaCl at pH 7.
Reproduced with permission.31
Reaction time. Although the incubation in Ab-BA 1, 2, 4, 6, 7, 10 and 11 was done for 12
h or more at 4 °C, it was shown in Ab-BA 8 that the Ab immobilization was complete
within 20 min.
Blocking. After Ab immobilization, the remaining surface binding sites (any sites that may
attract unwanted proteins in subsequent experiments, e.g., by boronic acid-carbohydrate,
hydrophobic or electrostatic interactions) were blocked with dextran (Ab-BA 1, 6 and 10),
bovine serum albumin (Ab-BA 3, 4, 5, 6, 8, 10 and 11), casein (Ab-BA 7), milk (Ab-BA 9)
or left unblocked (Ab-BA 2). In Ab-BA 1, five different compounds (glucose, glycerol,
dextran, ethylene glycol and ethanolamine) were tested for blocking. Dextran was selected
as it resulted in the most effective suppression of non-specific binding, although we expect
that it may expel the Ab from the surface upon binding with the boronic acid.
Characterization of the antibody-functionalized surface. In order to pinpoint causes of
possibly unsuccessful antigen binding, it is useful to figure out whether the Ab is bound to
the surface in the expected way. Although Ab-BA 1, 4, 5, 6, 7 and 11 reported experiments
Key steps towards the immobilization of antibodies using boronic acids
23
that indicated that the Abs were successfully immobilized on the surfaces, the orientation of
the Abs was directly assessed in Ab-BA 8 only: dual polarization interferometry showed
upon binding of the Ab a thickness increase of 8.44 ± 0.86 nm and a mass increment of
3.33 ± 0.40 ng/mm2. These values indicated that the Abs were attached to the surface via
their Fc chain (“end-on” conformation), based on Ab dimensions of 14.2 × 8.5 × 4.0 nm
and a previously reported mass increment of 3.70 ng/mm2 for an end-on oriented Ab. More
information about the techniques to study Ab orientation can be found in a previous review
by our group.4
2.4 Stability
The reaction between a boronic acid and a diol to form a boronate ester is reversible, which
may strongly disturb Ab immobilization. Diol content and pH of the solution both have a
strong influence on the equilibrium.
Effect of competing diols on the stability of the immobilized antibody. In Ab-BA 5 and
6, it was shown that sorbitol solutions were able to release the Ab from the surface. In Ab-
BA 1, a western blot (Figure S4 in Ab-BA 1) suggests a poor stability of the Ab-
functionalized material in serum. In order to prevent the competition of soluble diols with
the immobilized Ab, APBA was added to serum before incubation of the Ab-functionalized
material with this serum. Subsequent antigen binding was apparently improved, but we are
doubtful about the validity of this treatment: APBA may compete with the surface-bound
boronic acid and expel the Ab from the surface, resulting in a loss of antigen binding
capacity and consequently a wrong measurement of antigen concentration in the sample.
If the antigen is a protein, then chances are high that it is glycosylated.32 Therefore, it may
expel the Ab from the surface, resulting in erroneous measurements. Blocking antigen diols
with APBA may not only cause the problem described in the previous paragraph, but also
cause blocking of the antigen epitopes or nonspecific interactions between antigen and
surface. However, if the antigen is big enough and the Ab layer is dense enough, then steric
hindrance may prevent the antigen from accessing the boronic acid.33 If steric hindrance is
Chapter 2
24
not sufficient, then we recommend a synthetically involved but potentially effective
solution: make the boronic acid-mediated immobilization irreversible (see chapter 3).
Effect of pH on the stability of the immobilized antibody. Treatment with acidic
solutions, as reported in Ab-BA 6 (pH 3) and 8 (glycine pH 2.5), resulted in dissociation of
the Ab from the surface. In Ab-BA 1, a western blot (Figure S2 in Ab-BA 1) suggests an
increasing dissociation of the Ab from the surface when the pH was increased from 7 to 12.
If the immunoassay conditions stay in a “safe” pH range and the sensor surface is not
intended for reuse, then this should not cause a problem. If regeneration of the sensor
surface with an acidic or basic solution is required, then one must keep in mind that not
only the antigen, but also part or all of the Ab will be washed away, depending on the
regeneration conditions.
3 Conclusion and outlook
Ab immobilization using boronic acids requires substantial preliminary work: preparing a
boronic acid-containing antifouling surface (most likely via several chemical steps), finding
an Ab with suitable N-glycans, finding binding conditions that result in effective and
oriented Ab immobilization, and making sure that the immobilized Ab is not expelled from
the surface by competing diols or harsh regeneration conditions.
However, once the basic requirements are met and initial optimizations are done, Ab
immobilization via boronic acids may result in a much higher sensitivity for antigen
detection than what can be achieved by random immobilization, and an improvement of
cost and simplicity compared to other strategies for oriented Ab immobilization.
Future research in this field may give precious information about trends in the glycosylation
of Abs and their affinity for boronic acids, and explore the effect of boronic acid structure,
surface architecture and binding conditions on Ab immobilization and orientation. This
could lead to the development of Ab immobilization via boronic acids towards a well-
established method within the fabrication of high-performance biosensors. Moreover,
Key steps towards the immobilization of antibodies using boronic acids
25
further knowledge on the interactions between boronic acids and antibodies may lead to
improvements in related fields, such as the purification or site-specific labelling of
antibodies and, more generally, glycoproteins.
Reversibility of the binding between boronic acid and diol, which remains an obstacle of
Ab immobilization that relies only on this binding, has led to a strategy for oriented and
irreversible Ab immobilization. This strategy is described in chapter 3.
4 References
1. F. Rusmini, Z. Zhong and J. Feijen, Protein immobilization strategies for protein biochips,
Biomacromolecules, 2007, 8, 1775-1789.
2. P. Jonkheijm, D. Weinrich, H. Schroder, C. M. Niemeyer and H. Waldmann, Chemical strategies for
generating protein biochips, Angewandte Chemie-International Edition, 2008, 47, 9618-9647.
3. A. K. Trilling, M. M. Harmsen, V. J. B. Ruigrok, H. Zuilhof and J. Beekwilder, The effect of uniform
capture molecule orientation on biosensor sensitivity: Dependence on analyte properties, Biosensors
and Bioelectronics, 2013, 40, 219-226.
4. A. K. Trilling, J. Beekwilder and H. Zuilhof, Antibody orientation on biosensor surfaces: a minireview,
Analyst, 2013, 138, 1619-1627.
5. A. Makaraviciute and A. Ramanaviciene, Site-directed antibody immobilization techniques for
immunosensors, Biosensors and Bioelectronics, 2013, 50, 460-471.
6. X. Wang, N. Xia and L. Liu, Boronic acid-based approach for separation and immobilization of
glycoproteins and its application in sensing, International Journal of Molecular Sciences, 2013, 14,
20890-20912.
7. P. C. Lin, S. H. Chen, K. Y. Wang, M. L. Chen, A. K. Adak, J. R. R. Hwu, Y. J. Chen and C. C. Lin,
Fabrication of oriented antibody-conjugated magnetic nanoprobes and their immunoaffinity
application, Analytical Chemistry, 2009, 81, 8774-8782.
8. J. A. A. Ho, W. L. Hsu, W. C. Liao, J. K. Chiu, M. L. Chen, H. C. Chang and C. C. Li, Ultrasensitive
electrochemical detection of biotin using electrically addressable site-oriented antibody immobilization
approach via aminophenyl boronic acid, Biosensors and Bioelectronics, 2010, 26, 1021-1027.
9. M. Moreno-Guzmán, I. Ojeda, R. Villalonga, A. González-Cortés, P. Yáñez-Sedeño and J. M.
Pingarrón, Ultrasensitive detection of adrenocorticotropin hormone (ACTH) using disposable
phenylboronic-modified electrochemical immunosensors, Biosensors and Bioelectronics, 2012, 35, 82-
86.
Chapter 2
26
10. M. Moreno-Guzmán, A. González-Cortés, P. Yáñez-Sedeño and J. M. Pingarrón, Multiplexed
ultrasensitive determination of adrenocorticotropin and cortisol hormones at a dual electrochemical
immunosensor, Electroanalysis, 2012, 24, 1100-1108.
11. Z. Wang, K. Shang, J. Dong, Z. Cheng and S. Ai, Electrochemical immunoassay for subgroup J of
avian leukosis viruses using a glassy carbon electrode modified with a film of poly (3-thiophene
boronic acid), gold nanoparticles, graphene and immobilized antibody, Microchimica Acta, 2012, 179,
227-234.
12. L. Song, J. Zhao, S. Luan, J. Ma, J. Liu, X. Xu and J. Yin, Fabrication of a detection platform with
boronic-acid-containing zwitterionic polymer brush, ACS Applied Materials & Interfaces, 2013, 5,
13207-13215.
13. W. C. Liao and J. A. Annie Ho, Improved activity of immobilized antibody by paratope orientation
controller: Probing paratope orientation by electrochemical strategy and surface plasmon resonance
spectroscopy, Biosensors and Bioelectronics, 2014, 55, 32-38.
14. H. Y. Song, J. Hobley, X. Su and X. Zhou, End-on covalent antibody immobilization on dual
polarization interferometry sensor chip for enhanced immuno-sensing, Plasmonics, 2014, 9, 851-858.
15. M. González-González, R. Bartolome, R. Jara-Acevedo, J. Casado-Vela, N. Dasilva, S. Matarraz, J.
García, J. A. Alcazar, J. M. Sayagues, A. Orfao and M. Fuentes, Evaluation of homo- and hetero-
functionally activated glass surfaces for optimized antibody arrays, Analytical Biochemistry, 2014,
450, 37-45.
16. L.-D. Huang, A. K. Adak, C.-C. Yu, W.-C. Hsiao, H.-J. Lin, M.-L. Chen and C.-C. Lin, Fabrication of
highly stable glyco-gold nanoparticles and development of a glyco-gold nanoparticle-based oriented
immobilized antibody microarray for lectin (GOAL) assay, Chemistry – A European Journal, 2015, 21,
3956-3967.
17. S. A. Lim and M. U. Ahmed, A carbon nanofiber-based label free immunosensor for high sensitive
detection of recombinant bovine somatotropin, Biosensors and Bioelectronics, 2015, 70, 48-53.
18. J. A. Peters, Interactions between boric acid derivatives and saccharides in aqueous media: Structures
and stabilities of resulting esters, Coordination Chemistry Reviews, 2014, 268, 1-22.
19. H. Li and Z. Liu, Recent advances in monolithic column-based boronate-affinity chromatography,
TrAC Trends in Analytical Chemistry, 2012, 37, 148-161.
20. C. Rodriguez Emmenegger, E. Brynda, T. Riedel, Z. Sedlakova, M. Houska and A. B. Alles,
Interaction of blood plasma with antifouling surfaces, Langmuir, 2009, 25, 6328-6333.
21. S. Joshi, P. Pellacani, T. A. van Beek, H. Zuilhof and M. W. F. Nielen, Surface characterization and
antifouling properties of nanostructured gold chips for imaging surface plasmon resonance biosensing,
Sensors and Actuators B: Chemical, 2015, 209, 505-514.
22. D. Zhang, K. L. Thompson, R. Pelton and S. P. Armes, Controlling deposition and release of polyol-
stabilized latex on boronic acid-derivatized cellulose, Langmuir, 2010, 26, 17237-17241.
23. A. E. Ivanov, A. Kumar, S. Nilsang, M. R. Aguilar, L. I. Mikhalovska, I. N. Savina, L. Nilsson, I. G.
Scheblykin, M. V. Kuzimenkova and I. Y. Galaev, Evaluation of boronate-containing polymer brushes
Key steps towards the immobilization of antibodies using boronic acids
27
and gels as substrates for carbohydrate-mediated adhesion and cultivation of animal cells, Colloids
and Surfaces B: Biointerfaces, 2010, 75, 510-519.
24. P.-C. Chen, L.-S. Wan, B.-B. Ke and Z.-K. Xu, Honeycomb-patterned film segregated with
phenylboronic acid for glucose sensing, Langmuir, 2011, 27, 12597-12605.
25. R. Pelton, Y. Cui, D. Zhang, Y. Chen, K. L. Thompson, S. P. Armes and M. A. Brook, Facile
phenylboronate modification of silica by a silaneboronate, Langmuir, 2012, 29, 594-598.
26. Y.-W. Lu, C.-W. Chien, P.-C. Lin, L.-D. Huang, C.-Y. Chen, S.-W. Wu, C.-L. Han, K.-H. Khoo, C.-C.
Lin and Y.-J. Chen, BAD-lectins: Boronic acid-decorated lectins with enhanced binding affinity for the
selective enrichment of glycoproteins, Analytical Chemistry, 2013, 85, 8268-8276.
27. S. Raju, Glycosylation variations with expression systems, BioProcess International, 2003, 1, 44-53.
28. C. Huhn, M. H. J. Selman, L. R. Ruhaak, A. M. Deelder and M. Wuhrer, IgG glycosylation analysis,
Proteomics, 2009, 9, 882-913.
29. G. P. Roberts, Histochemical detection of sialic acid residues using periodate oxidation, Histochemical
Journal, 1977, 9, 97-102.
30. B. Zhang, S. Mathewson and H. Chen, Two-dimensional liquid chromatographic methods to examine
phenylboronate interactions with recombinant antibodies, Journal of Chromatography A, 2009, 1216,
5676-5686.
31. A. M. Azevedo, A. G. Gomes, L. Borlido, I. F. S. Santos, D. M. F. Prazeres and M. R. Aires-Barros,
Capture of human monoclonal antibodies from a clarified cell culture supernatant by phenyl boronate
chromatography, Journal of Molecular Recognition, 2010, 23, 569-576.
32. G. A. Khoury, R. C. Baliban and C. A. Floudas, Proteome-wide post-translational modification
statistics: Frequency analysis and curation of the swiss-prot database, Scientific Reports, 2014, 1.
33. S. Liu, X. Zhang, Y. Wu, Y. Tu and L. He, Prostate-specific antigen detection by using a reusable
amperometric immunosensor based on reversible binding and leasing of HRP-anti-PSA from
phenylboronic acid modified electrode, Clinica Chimica Acta, 2008, 395, 51-56.
Chapter 2
28
29
Chapter 3Design and synthesis of linkers for the oriented and irreversible immobilization of antibodies
To enable the oriented and irreversible immobilization of antibodies, two linkers were
designed with an amine for surface attachment, a boronic acid for capturing antibodies via
the N-glycans in their Fc chain, and a diazirine for irreversible immobilization upon UV
irradiation while maintaining antibody orientation. From a diazirine building-block that was
obtained in three steps, the first linker was synthesized in four steps and the second linker
was synthesized in three steps. Diol-functionalized silica was used for the chromatography
of two boronic acid-containing intermediates, this method being novel (to the best of our
knowledge) and likely based on boronic acid-diol interactions. High-resolution mass
spectrometry, through matching exact masses, matching isotope patterns and observation of
species corresponding to the esterification of boronic acids with MeOH, confirmed that
both linkers were synthesized successfully.
Florine Duval, Han Zuilhof and Teris A. van Beek are authors of this chapter.
Diol in N-glycanBoronic acid
Diazirine
Trifunctional linker
Antigen binding sites
Fc region
Antibody
Antibody immobilized on surface
Orientedimmobilization
Irreversibleimmobilization
Amine
Chapter 3
30
1 Introduction
The oriented immobilization of antibodies using boronic acids, as reviewed and discussed
in chapter 2, seems very promising. However, because the boronic acid-diol complexation
is reversible, it is difficult to apply this immobilization strategy to the fabrication of sensors
for the analysis of sugar-containing samples, or sensors that need to be regenerated using
acidic or basic solutions. When the surface is incubated with a sample to analyze or a
regeneration solution, the antibody may be washed away from the surface. Therefore, we
asked ourselves the following question: how to orient antibodies on surfaces using boronic
acids, and immobilize them in an irreversible way while maintaining their orientation?
2 Idea and design
2.1 Development of the idea
The idea of oriented and irreversible antibody immobilization was triggered by the article
of Abad et al. from 2002, which describes the immobilization of a peroxidase glycoprotein
on a mixed epoxy-boronic acid monolayer (Figure 1).1 As the diols in the glycoprotein
glycans react fast with the surface-bound boronic acid, the glycoprotein concentrates on the
surface. The surface-bound epoxide groups and the glycoprotein nucleophilic groups (such
as amines) come in close proximity, and are therefore able to react with each other in spite
of the slow kinetics of the reaction between an epoxide and a nucleophile. The formed bond
is very stable and the peroxidase glycoprotein is tightly bound to the surface.
Linkers for the oriented and irreversible immobilization of antibodies
31
Figure 1. Immobilization of peroxidase on mixed epoxy-boronic acid monolayer. 1 Reprinted with permission from 1.
Copyright 2015 American Chemical Society.
Can the same mixed epoxy-boronic acid monolayer directly be used for the oriented
immobilization of antibodies? No: first, the antibody N-glycans (see chapter 2) would bind
to the boronic acid on the surface. Second, the antibody would move around the obtained
anchor point and almost every part of the antibody would be able to access the epoxide
groups on the surface. This would result in a random immobilization of the antibody, or
maybe worse: the antibody sites that are away from the N-glycans, including the antigen
binding sites, may be even more susceptible to react with the epoxide groups.
Therefore, we thought of designing a surface bearing a long spacer, at the end of which a
boronic acid and an epoxide, close to each other, would be present. First, the boronic acid
would react with a diol in the antibody N-glycans. Second, the epoxide would react with a
nucleophile on the antibody in the vicinity of the N-glycans. At the end, regardless of the
reversibility of the binding between diol and boronic acid, the antibody would be bound to
the surface via its Fc chain in an irreversible way.
However, the use of an epoxide would present the following drawbacks: 1) The epoxide
should be formed just before antibody immobilization, on the sensing surface, and the
surface may not resist the harsh conditions (such as peracid or concentrated base) that are
necessary for the oxidation of an alkene (common stable precursor of epoxides) into an
epoxide. 2) Whereas most epoxides in a reaction mixture will react with a nucleophile after
the boronic acid reacts with a diol, some epoxide moieties will react before their
Chapter 3
32
neighboring boronic acid, resulting in improper orientation of the antibody. 3) The reaction
of epoxides with the nucleophilic groups present on proteins is slow. On the one hand, this
is what makes the approach possible. On the other hand, the reaction may therefore also be
inefficient as some epoxide groups may hydrolyze before they react with the antibody.
4) Enough nucleophiles should be present in the close proximity of the N-glycans for the
antibody to be immobilized in an irreversible way.
To circumvent all these problems together, we decided to replace the epoxide by a
photoreactive group. Photoreagents such as aryl azides, benzophenones and diazirines
attached to surfaces have been used for protein immobilization upon irradiation with UV
light.2 Diazirines, in particular, require a shorter irradiation time than benzophenones and a
higher irradiation wavelength than aryl azides, thereby minimizing protein degradation.
Upon irradiation with 365 nm light, diazirine groups generate carbenes, which rapidly
insert into C-C, C-H and X-H (X = heteroatom) bonds, thereby forming covalent bonds
with neighboring (bio)molecules (Figure 2, example with C-H bond).3 Trifluoromethyl-
aryldiazirines are among the most chemically and thermally stable of all types of diazirines
and the formed carbenes do not participate in intramolecular rearrangements. For these
reasons, trifluoromethylaryldiazirines are most commonly used.4
Figure 2. Chemistry of the photoimmobilization of (bio)molecules with diazirines.
In order to prepare a surface bearing a long spacer, at the end of which a boronic acid and a
diazirine would be present, we envisaged the synthesis of a trifunctional linker that would
contain 1) a primary amine for surface attachment, 2) a boronic acid and 3) a diazirine. The
obtained linker would then be attached to a surface bearing a long spacer (e.g., polyethylene
glycol) that is terminated with a reactive ester (e.g., N-hydroxysuccinimide ester). Such
surfaces are commercially available. The strategy for oriented and irreversible antibody
immobilization via boronic acid and diazirine is summarized in Figure 3.
Linkers for the oriented and irreversible immobilization of antibodies
33
Figure 3. Strategy for oriented and irreversible antibody immobilization via boronic acid and diazirine.
Boronic acid
DiazirineAmine
5) Upon irradiation with UVlight, diazirine is convertedinto a reactive carbene.
4) Boronic acid binds with diol, antigen binding sites arekept away from the surface.
6) Carbene reacts immediatelywith a C-C, C-H or X-H bond (X = heteroatom) in antibodybackbone.
7) If boronic acid breaks up with diol, antibody is still bound to the surface, bindingsites are still away from the surface.
3) Trifunctional linker-coatedsurface is incubated withantibody.
Long spacer
Diol in N-glycan
Sensor surface
Antibody
Antigen binding sites
Fc region
1) Activated ester-coated sensor surface isincubated with trifunctional linker.
Activated ester
2) Amine reacts with activatedester and trifunctional linker isattached to surface.
Trifunctional linker
Chapter 3
34
2.2 Linker design and synthetic approach
The described molecules are depicted in Figure 4.
Diazirine building-block. In order to introduce a trifluoromethylaryldiazirine in the
linkers, compound 1 was chosen as it is commercially available and contains a carboxylic
acid group that would allow straightforward attachment to an amine-containing molecule.
When the boronic acid is bound to the diol in the antibody N-glycan and the diazirine is
converted into a reactive carbene, the carbene should have enough freedom to reach the
antibody. To allow this, we chose to introduce a spacer within the trifunctional linker.
Diamine 2 would be connected to carboxylic acid 1 and the obtained diazirine building-
block 3 would present a diazirine on one side and a primary amine on the other side for
attachment to the rest of the linker.
Linker A. 3-Aminophenylboronic acid 4 (APBA) was chosen as the starting material for
the introduction of a boronic acid group in the linker. As discussed in chapter 2, other
boronic acids may result in a stronger interaction with diols, but APBA seemed a good
starting point for proof-of-principle experiments: it is by far the most commonly used
boronic acid for antibody immobilization (see chapter 2), and using the same structure
would allow an easier comparison of our strategy with known Ab immobilization via
boronic acids. For the linker backbone, glutamic acid 5 was chosen as it conveniently
contains an amine for attachment of the finished linker to the surface, a carboxylic acid for
reaction with the amine group of APBA and another carboxylic acid for reaction with the
amine group of the diazirine building-block 3.
Linker B. Having difficulties with the synthesis and purification of Linker A
intermediates, Linker B was designed from the need for a more straightforward synthetic
route towards a linker with the desired functional groups (amine, boronic acid and
diazirine). Addition reactions between thiols and terminal alkenes, namely thiol-ene “click”
reactions, are known to be very easy, versatile and efficient reactions, often directly
yielding pure products.5 Therefore, this type of reaction seemed attractive for the synthesis
of boronic acid-containing compounds, which are typically difficult to purify. The terminal
Linkers for the oriented and irreversible immobilization of antibodies
35
alkene of commercially available boronic acid 6 would “click” with the thiol of cysteine 7,
which also contains an amine for attachment of the finished linker to the surface and a
carboxylic acid for reaction with the amine group of the diazirine building-block 3.
Figure 4. Design and synthetic approach of diazirine building-block, Linker A and Linker B.
O
HN
O
CF3NN
O
H2N
O
NHO
BOH
OH
HN
H2N
O
CF3NN
O OHN
NH2
BOH
OH
H2N
O
OH
OHO
NH
O
HN
CF3N
O
N
O
O
HN
OH
OHB
S
H2NO
NH
O
H2N
CF3N
O
N
O
O
HN
OH
OHB
OH
SH
H2NO
Amideformation
Amideformation
Amideformation
Thiol-ene"click"
Linker A
Linker B
CF3NN
OHO NHO
O
CF3NN
OH2NNH2OOH2N
Amideformation
Diazirine building-block1
2 3
4
5 3
3
6
7
Boronic acid
DiazirineAmine
Chapter 3
36
3 Results and discussion
3.1 Synthesis of the diazirine building-block
It may be possible to directly couple diamine 2 with carboxylic acid 1 and obtain diazirine
building-block 3 in one step (Figure 5B). One molecule of diamine 2 can potentially react
with two molecules of carboxylic acid 1, but if a large excess of 2 is used, it may promote
the formation of 3 as a major compound. However, a mixture of diamine 2, desired product
3 and by-product 3a would be unavoidable. The delicate purification process would result
in a partial loss of compound 3, and thus a waste of expensive compound 1. To avoid this,
we chose to first protect diamine 2 on one side with a Boc group, then couple the obtained
product 8 with compound 1, and finally remove the Boc group (Figure 5A).
Boc protection encountered the same issue as described above, as a mixture of diamine 2,
desired product 8 and by-product 8a was obtained (Figure 5C). However, the reagent for
Boc protection di-tert-butyl dicarbonate (Boc2O) is cheap and widely available, unlike
compound 1. The reaction was performed at a relatively large scale, and a sufficient amount
of amine 8 was isolated (3.02 g, 59% yield). After coupling amine 8 with carboxylic acid 1,
purification of the obtained amide 9 by silica flash chromatography, and subsequent
removal of the Boc group in typical conditions, the HCl salt of diazirine building-block 3
was isolated in a quantitative yield over the last two steps. The effort of working in the dark
to prevent degradation of the diazirine moiety proved fruitful (although the synthesis has
not been carried out in full light for yield comparison, NMR analysis of compound 1 before
and after exposure to laboratory light for 2 h showed about 10% degradation of the
material, which had shown excellent stability in the dark for several days).
Linkers for the oriented and irreversible immobilization of antibodies
37
Figure 5. Synthesis of the diazirine building-block: A. Synthetic strategy as performed; B. Synthetic strategy in
one step, which was not performed as it results in an undesired mixture of compounds and waste of the diazirine
starting material; C. First step of the synthetic strategy as performed, showing the obtained by-products.
Boc2O
DCMRT
H2N
O
O
NHBoc
H2N
O
O
NH2
2 8 9 3
1 HN
O
O
NHBoc
CF3
N
N
OHN
O
O
NH2
CF3
N
N
O
CF3
N
N
OHO
60%
EDC.HClHOBtDIPEA
DCMRT
100%
HCl
Et2ORT
100%
HN
OO
H2N
O
NH2OO
H2N
HN
OO
NH
O
O
O
O
2
8
2
8a
+
+
Large excess
NH2OO
H2N
+
O
O
O
O
OBoc2O
Mixture
HN
OO
H2N
CF3
N N
OCF3
N N
O
HONH2O
OH2N
HN
OO
NH
CF3
N N
O
O
F3C
N N
2
3
2
3a
1+
+
Large excess
NH2OO
H2N
+
Activatingagent
Mixture
A
B
C
Chapter 3
38
3.2 Synthesis of linker A
The synthesis of Linker A, as depicted in Figure 6, was done in four steps.
1) Amide formation. The first step towards Linker A consisted of the coupling between
the amine of commercially available APBA 4 and the carboxylic acid of commercially
available protected glutamic acid 10, which was performed in DCM and in the presence of
EDC.HCl as activation reagent. Liquid-liquid extraction was not deemed suitable to isolate
product 11, as many boronic acid-containing compounds have an amphiphilic character.
This makes the extraction of such compounds from aqueous solutions difficult.6 Therefore,
the reaction mixture was directly purified by silica flash chromatography, and the desired
product 11 was obtained in 30% yield.
2) Carboxylic acid deprotection. Removal of the tert-butyl group from the remaining
carboxylic acid was achieved in TFA/DCM 1:1, typical conditions for this type of reaction.
It did not require any purification and afforded compound 12 in a 98% yield.
3) Amide formation. The coupling of amine-containing diazirine building-block 3 and
carboxylic acid 12 was performed in DCM, with a slight excess of 3, in presence of
EDC.HCl and DIPEA. TLC showed complete conversion of acid 12. After a failed attempt
to isolate compound 13 by silica flash chromatography, the reaction was repeated and the
desired product was successfully isolated by means of diol-silica flash chromatography (see
3.4). Compound 13 contained major impurities by 1H NMR, but it was used as such for the
next step in the hope that the final product would be easier to purify.
4) Amine deprotection. Removal of the Fmoc protecting group was performed in 20%
piperidine in DMF, typical conditions for this type of reaction. This is normally a relatively
clean reaction, as only two products are formed (plus release of CO2): the free amine and
the Fmoc by-product. In our case, however, 1H NMR showed that a complex mixture was
obtained, and too little material was available for any purification attempt. Electrospray
ionization high-resolution mass spectrometry (ESI-HRMS) (Figure 13) did indicate the
presence of the desired final compound, Linker A.
Linkers for the oriented and irreversible immobilization of antibodies
39
O
HN
O
CF3NN
O
H2N
O
NHO
B
OH
OH
HNO
HN
O
CF3NN
O
FmocHN
O
NHO
B
OH
OH
HN
3EDC.HClDIPEA
DCMRT
Product formed
O
OH
FmocHN
NHO
B
OH
OH
O
OtBu
FmocHN
NHO
B
OH
OH
O
OtBu
FmocHN
OHO NH2
B
OH
OH
Piperidine
DMF
Product formed
10
4
11 12
13 Linker A
EDC.HCl
DCMRT
TFA
DCMRT
98%30%
Figure 6. Synthesis of Linker A.
Chapter 3
40
3.3 Synthesis of linker B
The synthesis of Linker B, as depicted in Figure 7, was done in three steps.
1) Thiol-ene “click”. The first step consisted of a radical thiol-ene reaction between the
thiol of commercially available Boc-protected cysteine 14, and the alkene of commercially
available boronic acid 6. A large excess of 14 was used in order to push the desired reaction
to completion. DMPA was the radical initiator, and the mixture was irradiated with UV
(365 nm). This reaction was done four times:
1. At test scale (15 mg alkene), the reaction was monitored by TLC, which showed near-
complete conversion of the alkene after 40 min.
2. To ensure complete conversion of the alkene, the reaction was repeated (100 mg alkene)
and run for 2 h. Diol-silica flash chromatography (see 3.4) effectively removed DMPA and
excess thiol 14. The desired product 15 was isolated, but some impurities were present.
Subsequent purification by preparative reversed-phase HPLC resulted in a much higher
purity as shown by 1H NMR analysis.
3. The reaction was repeated at a larger scale (500 mg alkene). After two failed attempts to
purify the product by reversed-phase flash column chromatography (no separation
occurred), the recovered material was purified by diol-silica flash chromatography (see 3.4)
as in the previous reaction. When further purification was attempted by preparative
reversed-phase HPLC, analysis of the obtained fractions by ESI-HRMS did not show the
desired product. Instead, the data suggested that the boronic acid had oxidized to a phenol,
and the thio-ether had oxidized to a sulfoxide (see chapter 6 for further discussion).
4. The reaction was repeated under the same conditions (500 mg alkene), and compound 15was directly isolated by diol-silica flash chromatography (see 3.4). 1H NMR suggested the
presence of some impurities, but no further purification was attempted as it might result in
degradation of the product.
2) Amide formation. The coupling of carboxylic acid 15 with amine-containing diazirine
building-block 3 was performed in DCM, with a slight excess of 3, in presence of EDC.HCl
Linkers for the oriented and irreversible immobilization of antibodies
41
and DIPEA (same conditions as for the synthesis of Linker A). TLC (using the method
described in chapter 4) showed the presence of various boronic acid-containing compounds,
thus diol-silica flash chromatography did not seem to be a good option for isolation of the
desired product. Preparative TLC was tested, without success. Purification by preparative
reversed-phase HPLC, however, afforded a trace (unknown) amount of compound 16, as
suggested by 1H NMR and confirmed by ESI-HRMS. The material was not pure, but the
available amount did not allow any further purification attempts.
3) Amine deprotection. Removal of the Boc group from compound 16 was carried out
using HCl in Et2O and ESI-HRMS proved the presence of the desired final compound,
Linker B.
Figure 7. Synthesis of Linker B.
NH
O
HN
CF3N
O
N
O
O
HN
OH
OHB
S
H2NO
NH
O
HN
CF3N
O
N
O
O
HN
OH
OHB
S
BocHNO
HCl
Et2O
Product formed
OH
O
HN
OH
OHB
S
BocHNO
O
HN
OH
OHB
OH
SH
BocHNO
DMPA
THF365 nm light, RT
88%
15EDC.HClDIPEA
DCMRT
Product formed14
6
15
16 Linker B
Chapter 3
42
3.4 Diol-silica column chromatography of boronic acids
Boronic acids are difficult to purify by silica flash chromatography. In our own experience,
many compounds could not be recovered from the column, or were obtained in low yields
(such as compound 11). This results from the tendency of boronic acids to stick to silica
(probably by esterification with the silanols), even when highly polar eluents are used.6 To
circumvent this problem, we looked for an alternative way to purify the synthesized boronic
acid-containing compounds.
Several polystyrene (PS)-based resins have been developed for the capture and release of
boronic acids (see Table 1 for a brief overview), which is based on the pH/solvent-
dependent affinity between boronic acid and diol (or triol).
Ref. Authors Ligand Capture Release
7 Carboni et al. 1,3-diol Refluxing THF Refluxing MeOH, THF and DCM
8, 9 Hall et al. Diethanolamine THF THF/H2O/HOAc (90/5/5) or THF/H2O (9/1)
10 Yang et al. Catechol
11 Liu et al. Trometamine
DCM, THF or toluene THF/H2O/HOAc (90/5/5)
THF THF/H2O/HOAc (90/5/5)
Table 1. Polystyrene-based resins for the capture and release of boronic acids.
In particular, Hall et al. have developed a diethanolamine-functionalized PS resin (DEAM-
PS), which has become commercially available and enables efficient capture, derivatization
and release of boronic acids in mild conditions (Figure 8).8,9 The binding is not only based
on boronic acid-diol complexation, but also on an additional interaction between the
boronic acid (= Lewis acid) and the tertiary amine (= Lewis base) of PS-bound
diethanolamine.
Figure 8. Capture and release of boronic acids by diethanolamine-functionalized polystyrene resin (DEAM-PS)
RB
OH
OH
Dry THF
5% H2Oin THF Boronic acid captured by DEAM-PSBoronic acidDEAM-PS
Linkers for the oriented and irreversible immobilization of antibodies
43
We prepared a small amount of DEAM-PS using Hall’s procedure. Briefly, chloromethyl
PS resin (Merrifield resin) was treated with an excess of diethanolamine, resulting in
nucleophilic substitution of the chloride with the amine. The obtained material was not
characterized, but it was tested for boronic acid capture and release. In our hands, the
capture of APBA in dry THF and its release in 5% water in THF resulted in about 50%
APBA recovery.
In parallel, we got acquainted with commercially available diol-functionalized silica (diol-
silica, Figure 9). Silica powder seemed easier to work with than PS resin: it does not (have
to) swell, so it is compatible with many solvents, it does not break under mechanical stress
so mixtures can be vigorously stirred, and it does not stick to glassware. According to the
manufacturer, diol-silica is meant for the scavenging of boronic acids (e.g., excess boronic
acid in a Suzuki reaction), and subsequent release of the boronic acids from the material
would require harsh acidic or basic conditions. However, test capture and release of APBA
with diol-silica under the same conditions as with DEAM-PS gave similar results as
DEAM-PS, although the binding mode was different (no amine). Furthermore, diol-silica
was added to the water-free mixture resulting from the reaction of boronic acid-containing
intermediate 12 with an excess of diazirine building-block 3, reasoning that the obtained
boronic acid 13 should be captured selectively (if it was the only boronic acid present in
solution). After separating the diol-silica from the solution and washing it with dry THF,
the bound molecules were released using 5% water in THF. TLC of the obtained material
suggested the presence of boronic acid 13, although it was impure.
Figure 9. Capture and release of boronic acids by diol-functionalized silica.
Diol-silica is well known as a versatile stationary phase in chromatography, as it is neutral,
moderately polar and able to form hydrogen bonds with solvents and analytes. It has been
used in normal-phase, size-exclusion,12 hydrophilic interaction and reversed-phase
chromatography13 for the separation of diverse compounds. In particular, diol-silica has
Chapter 3
44
sometimes been used in “manual” column chromatography for the purification of organic
compounds.14,15
The retention of boronic acids on a diol-silica column has once been reported in 1992.16
The effect of pH of the mobile phase (consisting of 0.1 M phosphate solution of which pH
was varied from 4 to 9) and of the boronic acid structure on the retention time was studied.
Although the results did not lead to general conclusions, some of them suggested that
boronic acids with lower pKa values had higher retention times, and that ortho-substituted
phenylboronic acids could not be retained, most probably because of steric hindrance. This
paper has never been cited, and no further work in this direction could be found in the
literature.
As the capture and release of our boronic acid-containing intermediate 13 by diol-silica
looked promising, we tested diol-silica as stationary phase for the flash chromatography of
this compound. Although product and impurities ran very close to each other on silica TLC,
the diol-silica column enabled the rapid elution of these impurities in a gradient of MeOH
in EtOAc, and then the slow elution of boronic acid 13 when 1% acetic acid was added to
the eluent. The obtained material still contained major impurities, but the fact that some
material was recovered from the column was already an improvement compared to
previously attempted silica flash chromatography. Diol-silica flash chromatography was
also tested for the purification of boronic acid 15, which proved particularly effective
(Figure 10). According to silica TLC, boronic acid 15 and an apparently boronic acid-free
impurity had similar polarities, and Boc-cysteine 14, which was still present in large excess
in the reaction mixture, was difficult to remove by silica flash chromatography. However,
14, DMPA and the unknown impurity were quickly eluted from the diol-silica column,
whereas boronic acid 15 was eluted at a later stage and was obtained in a satisfying purity
according to 1H NMR. Diol-silica, therefore, seems a promising alternative to bare silica for
the column chromatography of boronic acid-containing compounds.
Linkers for the oriented and irreversible immobilization of antibodies
45
Figure 10. Purification of boronic acid-containing intermediate 15 by diol-silica column chromatography. Silica
TLC of the column fractions (10% MeOH and 1% acetic acid in EtOAc). Spots circled by pencil: visible under
254 nm light.
Reactionmixture
Boc-Cys 14
DMPA
DMPA
Boronic acid-containing product 15
Amide-containing product 15
Unknown impurity(boronic acid-free)
Unknown impurity
Reactionmixture
Alizarin staining: boronic acids (method described in chapter 4)
Ninhydrin staining: amino acids
Chapter 3
46
3.5 Mass spectrometry of boronic acids
All synthesized boronic acid-containing compounds were dissolved in MeOH and analyzed
by ESI-HRMS, and the obtained mass spectra presented a particular feature: species
corresponding to the reaction of the compounds with one or two molecules of MeOH were
identified (Figure 11). This is supported by the work of Wang et al. who studied the
chemistry of boronic acids under electrospray conditions and observed the same
phenomenon.17 We found this very useful because it proved, in addition to the matching
exact masses, the presence of boronic acid moieties with great confidence. All observed
masses are given in paragraph 4.5 and mass spectra of Linker A and Linker B are given in
Figure 13 and Figure 14, respectively. Moreover, as shown in Figure 12, the main species
observed in the analysis of both linkers in positive mode and negative mode revealed
isotope patterns that were in excellent agreement with the simulated ones.
Figure 11. Reaction of boronic acids with MeOH as observed in mass spectra.
Linkers for the oriented and irreversible immobilization of antibodies
47
Figure 12. Measured and simulated isotope patterns of the largest quasi-molecular ions observed in the HRMS
analysis of linkers A and B.
716 717 718 719 720 721 722m/z
0
20
40
60
80
100 717.225
719.222718.228716.229
720.225721.229
704 705 706 707 708 709m/z
0
20
40
60
80
100 705.246
706.249704.250
707.253 708.245
656 657 658 659 660 661m/z
0
20
40
60
80
100 657.222
659.219658.225656.225
660.222661.225
622.0 622.5 623.0 623.5 624.0 624.5 625.0 625.5 626.0m/z
0
20
40
60
80
100 623.261
624.264622.264
625.267
Positive mode: C27H35O7N6BF3 (M + CH2 + H+) Negative mode: C27H34O7N6BClF3 (M + CH2 + Cl–)
0
20
40
60
80
100
Rel
ativ
e Ab
unda
nce
657.224
659.222
658.227656.227
660.224661.218658.474
0
20
40
60
80
100
Rel
ativ
e Ab
unda
nce
717.227
719.224718.230
716.231
720.227721.221
717 2250
20
40
60
80
100
Rel
ativ
e Ab
unda
nce
705.247
706.249
704.250
707.252
705 246
0
20
40
60
80
100
Rel
ativ
e Ab
unda
nce
623.260
624.262622.263
625.265
623 261
622.263
623.260
624.262
625.265656.227
657.224
658.227659.222
660.224 661.218
622.264 624.264
625.267
656.225
657.222
658.225 659.219
660.222661.225
704.250
705.247
706.249
707.252
704.250
705.246
706.249
707.253 708.245
716.231718.230
717.227
719.224
720.227 721.221
716.229 718.228
717.225
719.222
720.225721.229
658.474
623.261
Linker A
Linker B
Measured
Simulated
Measured
Simulated
Measured
Simulated
Measured
Simulated
Positive mode: C29H38O7N6BF3NaS (M + CH2 + Na+) Negative mode: C29H38O7N6BClF3S (M + CH2 + Cl–)
Chapter 3
48
Figure 13. High-resolution mass spectra of Linker A.
600 610 620 630 640 650 660 670 680 690 700m/z
0
10
20
30
40
50
60
70
80
90
100
Rel
ativ
e A
bund
ance
623.2596C27 H35 O7 N6 B F3
-1.1355 mmu
645.2410C27 H34 O7 N6 B F3 Na
-1.6514 mmu
609.2442C26 H33 O7 N6 B F3
-0.7991 mmu
600 610 620 630 640 650 660 670 680 690 700m/z
0
10
20
30
40
50
60
70
80
90
100
Rel
ativ
e A
bund
ance
657.2242C27 H34 O7 N6 B Cl F3
2.4613 mmu
635.2629C28 H35 O7 N6 B F3
2.1896 mmu 671.2394C28 H36 O7 N6 B Cl F3
2.0049 mmu621.2471C27 H33 O7 N6 B F3
2.0652 mmu
M + CH2 + H+
M + H+
M + 2 CH2 – H+
M + CH2 + Cl–
M + 2 CH2 + Cl–
Positive mode
Negative mode
608.2378C26H32O7N6BF3
Calculated M
Linker A
M + CH2 – H+
M + Na+
M + CH2 + Na+
631.2245C26 H32 O7 N6 B F3 Na
-2.5214 mmu
659.2559C28 H36 O7 N6 B F3 Na
-2.3975 mmu
M + 2 CH2 + Na+
(trace) 645.2410
C27 H34 O7 N6 B F3 Na-1.6514 mmu
Linkers for the oriented and irreversible immobilization of antibodies
49
Figure 14. High-resolution mass spectra of Linker B.
660 670 680 690 700 710 720 730 740 750 760m/z
0
10
20
30
40
50
60
70
80
90
100
Rel
ativ
e A
bund
ance
705.2468C29 H38 O7 N6 B F3 Na S
0.8217 mmu
677.2410
719.2623C30 H40 O7 N6 B F3 Na S
0.6136 mmu
691.2314C28 H36 O7 N6 B F3 Na S
1.0909 mmu
663.2249 708.2437 749.2715722.2578 735.0274
674.8908
660 670 680 690 700 710 720 730 740 750 760m/z
0
10
20
30
40
50
60
70
80
90
100
Rel
ativ
e A
bund
ance
717.2276C29 H38 O7 N6 B Cl F3 S
2.5043 mmu
681.2505C29 H37 O7 N6 B F3 S
2.0928 mmu731.2429
C30 H40 O7 N6 B Cl F3 S2.2009 mmu
683.2655C29 H39 O7 N6 B F3 S
1.4468 mmu
M + CH2 – H+
M + 2 CH2 – H+
M + CH2 + Cl –
M + 2 CH2 + Cl–
668.2411C28H36O7N6BF3S
Calculated M
M + CH2 + Na+
M + CH2 + H+
M + Na+
M + 2 CH2 + H+
M + 2 CH2 + Na+
Positive mode
Negative mode
Linker B
697.2808C30 H41 O7 N6 B F3 S
1.0519 mmu
667.2347C28 H35 O7 N6 B F3 S
1.9088 mmu
695.2661C30 H39 O7 N6 B F3 S
2.0600 mmu
M – H+703.2113
C28 H36 O7 N6 B Cl F3 S1.8737 mmu
M + Cl–
Chapter 3
50
4 Experimental section
4.1 General information
All solvents and reagents were obtained commercially and used as received unless stated
otherwise. Moisture-sensitive reactions were performed under Ar atmosphere and with
oven-dried flasks. All experiments involving compounds that contain a diazirine group
were carried out at RT or lower temperature and in the dark or with minimal light. Diol-
functionalized silica was obtained from Screening Devices.
Reactants and reagents. 2,2′-(ethylenedioxy)diethylamine, HOBt, DIPEA, 3-
aminophenylboronic acid (APBA), Fmoc-Glu(OtBu)-OH, piperidine and DMPA were from
Sigma-Aldrich. Di-tert-butyl dicarbonate (Boc2O) and TFA were from Acros Organics. 4-
[3-(trifluoromethyl)-3H-diazirin-3-yl]benzoic acid was from TCI Europe. EDC.HCl was
from Fischer. 4-Allylaminocarbonyl)phenylboronic acid and N-alpha-tert-
butyloxycarbonyl-L-cysteine were from ABCR. HCl 1 M in Et2O was from Alfa-Aesar.
Solvents. For reactions, DCM and THF, both from Sigma-Aldrich, were purified using a
Pure Solv 400 solvent purification system (Innovative Technology, Amesbury, USA). DMF
was from VWR. DCM and EtOAc from Acros Organics, acetic acid and heptane from
VWR, MeOH and methyl tert-butyl ether from Biosolve and ammonia (25% in water) from
Merck were used for purifications. CDCl3 from Acros Organics and CD3OD from Roth
were used for NMR analysis.
Apparatus. 1H NMR spectra were recorded on a Bruker Avance III 400 MHz
spectrometer. Chemical shifts are reported in parts per million (ppm), with
tetramethylsilane as internal standard. High-resolution mass spectrometry analyses were
recorded on a Thermo Fisher Scientific Exactive High-Resolution mass spectrometer, using
electrospray ionization. The preparative HPLC setup consisted of two LC-8A pumps, SIL-
10AP auto-injector, FRC-10A fraction collector and SPD-M10Avp diode array detector
from Shimadzu, and an Alltima C18 column (250 × 22 mm, 5 µm) from Alltech.
Linkers for the oriented and irreversible immobilization of antibodies
51
4.2 Diazirine building-block
8. A solution of 2,2′-(ethylenedioxy)diethylamine 2 (15.5 mL,
106.2 mmol) in dry DCM (200 mL) was cooled to 0 °C. A solution of di-tert-butyl
dicarbonate (Boc2O, 4.52 g, 20.7 mmol) in dry DCM (50 mL) was added over 40 min. The
reaction mixture was allowed to warm up to RT and stirred overnight at RT. The reaction
mixture was purified by flash chromatography (silica, 1% to 20% MeOH/ammonia 9:1 in
DCM), to give 3.02 g of compound 8 (59%) and 1.55 g of a mixture of compounds 8 and
8a. Compound 8: 1H NMR (400 MHz, CDCl3) δ 1.40 (s, 9H), 2.84 (m, 2H), 3.27 (m, 2H),
3.50 (m, 4H), 3.58 (apparent s, 4H).
9. A solution of 4-[3-(trifluoromethyl)-3H-diazirin-3-
yl]benzoic acid 1 (500 mg, 2.17 mmol, 1 eq), 8 (701 mg,
2.82 mmol, 1.3 eq), HOBt (440 mg, 3.26 mmol on
anhydrous basis, 1.5 eq) and DIPEA (1.14 mL, 6.52 mmol, 3 eq) in dry DCM was cooled to
0 °C. EDC.HCl (625 mg, 3.26 mmol, 1.5 eq) was added, the reaction mixture was allowed
to warm up to RT and stirred overnight. TLC showed complete conversion of 8. The
reaction mixture was concentrated under vacuum, and the material was purified by flash
chromatography (30 g silica, 30% to 70% EtOAc in heptane) to give 1.04 g yellow oil
(“104%”, probably near-quantitative yield and unknown impurities). 1H NMR (400 MHz,
CD3OD) δ 1.37 (s, 9H), 3.15 (m, 2H), 3.45 (t, 2H), 3.50-3.64 (m, 8H), 7.30 (d, 2H), 7.87 (d,
2H).
3. Compound 9 (1.04 g, 2.17 mmol + impurities) was
dissolved in HCl 1 M in Et2O (30 mL), and stirred at RT
for 48 h. TLC (EtOAc) showed complete conversion of 9(incomplete after 24 h). The reaction mixture consisted of a white suspension and a wax on
the bottom of the flask. NMR analysis showed that the desired product was only in the wax.
The liquid was removed, the wax was dissolved in 2 mL MeOH, 10 mL HCl 1 M in Et2O
was added, which caused precipitation, and this mixture was stirred for 1 h (to make sure
that the product was only in HCl salt form). The mixture was concentrated under vacuum to
HN
OO
BocHN
CF3
N N
O
Chapter 3
52
give 890 mg waxy solid (quantitative yield). 1H NMR (400 MHz, CD3OD) δ 3.04 (m, 2H),
3.54 (m, 2H), 3.64 (m, 8H), 7.30 (d, 2H), 7.87 (d, 2H).
4.3 Linker A
11. A solution of APBA 4 (259 mg, 1.89 mmol, 1 eq) and Fmoc-
Glu(OtBu)-OH 10 (965 mg, 2.27 mmol, 1.2 eq) in dry DCM was cooled to
0 °C. EDC.HCl (905 mg, 4.72 mmol, 2.5 eq) was slowly added, and the
reaction mixture was stirred for five days (a shorter time may have been
sufficient). The reaction mixture was concentrated under vacuum and
purified by flash chromatography (silica, 50% to 100% EtOAc in heptane, and 5% MeOH
in EtOAc), to give 309 mg solid (30%). 1H NMR (400 MHz, CD3OD) δ 1.43 (s, 9H), 1.9
(m, 1H), 2.1 (m, 1H), 2.36 (t, 2H), 4.20 (t, 1H), 4.27 (m, 1H), 4.39 (m, 2H), 7.00-8.00
(several signals, 12H).
12. Compound 11 (112 mg, 0.206 mmol) was dissolved in DCM (3 mL).
TFA (3 mL) was added and the reaction mixture was stirred at RT
overnight. TLC (5% MeOH and 1% acetic acid in EtOAc) showed
complete conversion of 11 and a single new spot by UV and by alizarin
staining (method described in chapter 4). The reaction mixture was
concentrated under vacuum to give 98 mg light-brown solid (98%). 1H NMR (400 MHz,
CDCl3) δ 2.02 (m, 1H), 2.17 (m, 1H), 2.46 (t, 2H), 4.15-4.35 (m, 2H), 4.41 (m, 2H), 7.10-
7.90 (multiple signals, 12H).
13. A solution of 12 (38 mg, 0.078 mmol), 3 (41 mg, 0.10
mmol, 1.3 eq) and DIPEA (34 μL, 0.2 mmol, 2 eq) in dry
DCM (5 mL) was cooled to 0 °C. EDC.HCl (30 mg, 0.16
mmol, 2.0 eq) was slowly added, the reaction mixture was
allowed to warm up to RT and stirred overnight at RT. TLC
(10% MeOH and 1% acetic acid in EtOAc) showed complete
conversion of 12. The RM was concentrated from 5 mL to 2-3 mL and purified by flash
chromatography (8.4 g diol-silica, gradient of 0 to 5% MeOH in EtOAc and then 10%
O
OtBu
FmocHN
NHO
B
OH
OH
O
OH
FmocHN
NHO
B
OH
OH
O
HN
O
CF3NN
O
FmocHN
O
NHO
B
OH
OH
HN
Linkers for the oriented and irreversible immobilization of antibodies
53
MeOH and 1% acetic acid in EtOAc). The fractions that looked purest by TLC were
concentrated under vacuum at RT to give about 40 mg oil. 1H NMR suggested but did not
confirm that the desired product was present. 1H NMR (400 MHz, CDCl3) δ 1.94 (acetic
acid + 1H, based on chemical shift in Fmoc-Glu(OH)-APBA), 2.10 (m, 1H), 2.28 (t, 2H),
3.10-3.80 (several signals, 12H + MeOH + impurities), 4.10-4.25 (several signals, 1H +
1H), 4.33 (m, 1H), 6.40-8.00 (several signals, 16H), major unknown impurities.
Linker A. A solution of 13 in DMF (0.8 mL) and piperidine
(0.2 mL) was stirred at RT. When TLC (10% MeOH and 1%
acetic acid in EtOAc) showed complete conversion of 13, DMF
and piperidine were evaporated using a high-vacuum pump
(pressure between 0 and 1 mbar) with a liquid nitrogen trap. The
mixture (oil and solid) was partitioned between methyl tert-butyl
ether (6 mL) and water (6 mL). Not all material was dissolved, so some ammonia was
added: the water layer reached a pH of about 10-11. Both layers separated very slowly.
According to TLC analysis, both layers contained boronic acid-containing compounds, and
only the organic layer contained Fmoc by-product. The water layer was concentrated under
vacuum. NMR analysis of the obtained material was inconclusive, but ESI-HRMS
confirmed the presence of the desired final compound.
4.4 Linker B
15. (4-Allylaminocarbonyl)phenylboronic acid 6 (500 mg, 2.44
mmol, 1 eq), N-alpha-tert-butyloxycarbonyl-L-cysteine 14 (2.70
g, 12.2 mmol, 5 eq) and DMPA (625 mg, 2.44 mmol, 1 eq) were
dissolved in dry and stabilizer-free THF (10 mL). The resulting
solution was stirred and irradiated at 365 nm for 2 h. TLC (10% MeOH and 1% acetic acid
in EtOAc) showed complete conversion of 6. The reaction mixture was concentrated to a
volume of about 5 mL, and purified by flash chromatography (28 g diol-silica, 0 to 20%
MeOH in EtOAc) to give 920 mg solid/oil (88%). The obtained material was not very pure,
but it was used as such in the next step. Previous attempts to purify this compound by
preparative HPLC had resulted in its complete degradation. 1H NMR (400 MHz, CD3OD)
Chapter 3
54
δ 1.34 (s, 9H), 1.81 (quint, 2H), 2.57 (t, 2H), 2.76 (dd, 1H), 2.92 (dd, 1H), 3.38 (m, 2H),
4.11 and 4.22 (m, 1H), 7.50-7.80 (m, 4H).
16. A solution of 15 (100 mg, 0.23 mmol, 1 eq), HCl salt of 3(112 mg, 2.81 mmol, 1.2 eq) and DIPEA (0.16 mL, 0.94 mmol,
4 eq) in dry DCM was cooled to 0 °C. EDC.HCl (90 mg, 0.47
mmol, 2 eq) was added, the reaction mixture was allowed to
warm up to RT and stirred overnight at RT. Extensive TLC
analysis (various eluents, normal phase and reverse phase, two-
dimensional, analysis of spots by mass spectrometry) suggested that several spots contained
the desired product (possibly esterified with MeOH). Test preparative TLC and filtration
with reversed-phase silica were not successful. A part of the reaction mixture was purified
by preparative HPLC (reversed-phase) to give a few mg of a white solid. 1H NMR (400
MHz, CD3OD) δ 1.33 (s, 9H), 1.79 (q, 2H), 2.54 (t, 2H), 2.64 (dd, 1H), 2.82 (dd, 1H), 3.23-
3.60 (m, 14H + impurities), 4.12 (m, 1H), 7.23 (d, 2H), impurities.
Linker B. To 16 (unknown amount) was added 1 M HCl in Et2O
and the material did not dissolve. Some MeOH was added and the
material dissolved. The solution was stirred at RT overnight. The
reaction mixture, as analyzed by ESI-HRMS, contained the
desired final compound.
NH
O
HN
CF3N
O
N
O
O
HN
OH
OHB
S
BocHNO
Linkers for the oriented and irreversible immobilization of antibodies
55
4.5 High-resolution mass spectrometry
Compound 12 Compound 13 Linker A Calculated Measured Calculated Measured Calculated Measured
M 488.1755 830.3058 608.2378 Positive mode M + H+ 489.1828 831.3131 609.2450 609.2442 M + Na+ 511.1642 853.2945 631.2264 631.2245 M + CH2 + H+ 503.1984 845.3288 623.2607 623.2596 M + 2CH2 + H+ 517.2141 859.3444 637.2763
M + CH2 + Na+ 525.1798 525.1794 867.3102 645.2421 645.2410 M + 2CH2 + Na+ 539.1955 539.1957 881.3258 659.2577 659.2559
Negative mode M - H+ 487.1682 829.2986 607.2305 M + Cl- 523.1449 865.2752 643.2072
M + CH2 - H+ 501.1839 501.1827 843.3142 621.2461 621.2471 M + 2CH2 - H+ 515.1995 515.1982 857.3299 635.2618 635.2629 M + CH2 + Cl- 537.1605 879.2909 879.2925 657.2228 657.2242 M + 2CH2 + Cl- 551.1762 893.3065 893.3077 671.2385 671.2394
Compound 15 Compound 16 Linker B Calculated Measured Calculated Measured Calculated Measured
Positive mode M + H+ 427.1705 769.3008 669.2484 M + Na+ 449.1519 449.1523 791.2822 691.2298 691.2314 M + CH2 + H+ 441.1861 783.3165 683.2641 683.2655 M + 2CH2 + H+ 455.2018 797.3321 697.2797 697.2808 M + CH2 + Na+ 463.1675 463.1679 805.2979 805.2955 705.2455 705.2468 M + 2CH2 + Na+ 477.1832 819.3135 819.3117 719.2611 719.2623
Negative mode M - H+ 425.1559 425.1554 767.2863 667.2339 667.2347 M + Cl- 461.1326 803.2630 703.2105 703.2113 M + CH2 - H+ 439.1716 439.1720 781.3019 781.3022 681.2495 681.2505 M + 2CH2 - H+ 453.1872 453.1876 795.3176 795.3181 695.2652 695.2661 M + CH2 + Cl- 475.1483 817.2786 817.2795 717.2262 717.2276 M + 2CH2 + Cl- 489.1639 831.2943 831.2950 731.2418 731.2429 Compound 11: no recognizable fragment was found in its mass spectra
Chapter 3
56
5 Conclusion
Two linkers for the oriented and irreversible immobilization of antibodies were
synthesized. They contained an amine group to react with an active ester on a surface, a
boronic acid group to bind an antibody via its N-glycans in its Fc region, and a diazirine
group to irreversibly stick the antibody to the surface while maintaining its orientation.
However, when the final compounds were obtained as confirmed by high-resolution mass
spectrometry, Adak et al. published an article describing a linker with the exact same
functional groups (amine, boronic acid and trifluoromethylaryldiazirine), and they showed
that an antibody that was immobilized on a surface via this linker had a superior antigen-
binding ability.18 Our work was therefore no longer publishable, as the antibody
immobilization strategy that we had in mind had lost its novelty. Nevertheless, we would
like to highlight two points from this chapter that may be useful in future research: 1) a
novel building block, which consisted of a trifluoromethylaryldiazirine and a primary
amine connected to each other by a flexible chain, was efficiently synthesized and purified,
and our procedure may be used for other applications that require the attachment of a
photoreactive and flexible molecule to an electrophile-containing scaffold; 2) boronic acid-
containing intermediates were successfully isolated by means of diol-silica flash
chromatography. On the basis of this we propose the use of diol-silica as an alternative to
bare silica for the purification of boronic acids.
6 References
1. J. M. Abad, M. Vélez, C. Santamaría, J. M. Guisán, P. R. Matheus, L. Vázquez, I. Gazaryan, L.
Gorton, T. Gibson and V. M. Fernández, Immobilization of peroxidase glycoprotein on gold electrodes
modified with mixed epoxy-boronic acid monolayers, Journal of the American Chemical Society, 2002,
124, 12845-12853.
2. P. Jonkheijm, D. Weinrich, H. Schroder, C. M. Niemeyer and H. Waldmann, Chemical strategies for
generating protein biochips, Angewandte Chemie-International Edition, 2008, 47, 9618-9647.
3. L. Dubinsky, B. P. Krom and M. M. Meijler, Diazirine based photoaffinity labeling, Bioorganic &
Medicinal Chemistry, 2012, 20, 554-570.
Linkers for the oriented and irreversible immobilization of antibodies
57
4. A. Blencowe and W. Hayes, Development and application of diazirines in biological and synthetic
macromolecular systems, Soft Matter, 2005, 1, 178-205.
5. C. E. Hoyle and C. N. Bowman, Thiol–ene click chemistry, Angewandte Chemie International Edition,
2010, 49, 1540-1573.
6. D. G. Hall, Structure, properties, and preparation of boronic acid derivatives. Overview of their
reactions and applications, John Wiley & Sons: Weinheim, Germany, 2006.
7. B. Carboni, C. Pourbaix, F. Carreaux, H. Deleuze and B. Maillard, Boronic ester as a linker system for
solid phase synthesis, Tetrahedron Letters, 1999, 40, 7979-7983.
8. D. G. Hall, J. Tailor and M. Gravel, N,N-diethanolaminomethyl polystyrene: An efficient solid support
to immobilize boronic acids, 1999, 38, 3064-3067.
9. M. Gravel, K. A. Thompson, M. Zak, C. Bérubé and D. G. Hall, Universal solid-phase approach for
the immobilization, derivatization, and resin-to-resin transfer reactions of boronic acids, Journal of
Organic Chemistry, 2002, 67, 3-15.
10. W. Yang, X. Gao, G. Springsteen and B. Wang, Catechol pendant polystyrene for solid-phase
synthesis, Tetrahedron Letters, 2002, 43, 6339-6342.
11. F. L. Liu, S. B. Wan and T. Jiang, First preparation of a novel polyol resin for purifying arylboronic
acids, Chinese Chemical Letters, 2012, 23, 993-995.
12. C. F. Poole and S. K. Poole, Chromatography today, Elsevier, 1991.
13. P. Jandera, Stationary and mobile phases in hydrophilic interaction chromatography: a review,
Analytica Chimica Acta, 2011, 692, 1-25.
14. T. E. Smith, M. Djang, A. J. Velander, C. W. Downey, K. A. Carroll and S. van Alphen, Versatile
asymmetric synthesis of the kavalactones: first synthesis of (+)-kavain, Organic Letters, 2004, 6, 2317-
2320.
15. V. Damodaran, J. L. Ryan and R. A. Keyzers, Cyclic 3-alkyl pyridinium alkaloid monomers from a
New Zealand Haliclona sp. marine sponge, Journal of Natural Products, 2013, 76, 1997-2001.
16. G. A. Bektenova, V. V. Vekolova, P. P. Gladyshev and A. N. Karazhanova, Interactions of boronic
acids with diol-containing solid phases. II. Liquid-chromatographic study of interactions with diols,
Izvestiya Akademii Nauk Respubliki Kazakhstan, Seriya Khimicheskaya, 1992, 6, 69-74.
17. L. Wang, C. Dai, S. K. Burroughs, S. L. Wang and B. Wang, Arylboronic acid chemistry under
electrospray conditions, Chemistry – A European Journal, 2013, 19, 7587-7594.
18. A. K. Adak, B.-Y. Li, L.-D. Huang, T.-W. Lin, T.-C. Chang, K. C. Hwang and C.-C. Lin, Fabrication
of antibody microarrays by light-induced covalent and oriented immobilization, ACS Applied
Materials & Interfaces, 2014, 6, 10452-10460.
Chapter 3
58
59
Chapter 4Sensitive thin-layer chromatography detection of boronic acids using alizarin
A new method for the selective and sensitive detection of boronic acids on thin-layer
chromatography plates is described. The plate is briefly dipped in an alizarin solution,
allowed to dry in ambient air, and observed under 365 nm light. Alizarin emits a bright
yellow fluorescence only in the presence of a boronic acid.
This chapter is a slightly modified version of the following publication: F. Duval, T.A. van Beek, H. Zuilhof, Sensitive thin-layer chromatography detection of boronic acids using alizarin, Synlett, 2012, 23, 1751-1754.
O
O
OBO R
OH
TLC365 nmAlizarin
non-fluorescent fluorescent
Alizarin + boronic acid
0 1 2 3 5
hours of Suzuki reaction
Chapter 4
60
1 Introduction
Boronic acids are used in a wide range of applications (see chapter 1). Before any use, these
compounds first need to be synthesized. Easy detection of boronic acids during reactions,
however, remains a challenge.
The conversion of boronic acids has been monitored in Suzuki-Miyaura reaction mixtures
to which dihydroxycoumarins were added. Fluorescent complexes resulting from binding of
these compounds with boronic acids were visualized under a hand-held UV lamp (365
nm).1 TLC, however, is a more generally applicable and easier way to analyze a reaction
mixture. The detection of boronic acids on TLC plates would enable the chemist to monitor
not only the conversion of a boronic acid, but also the integrity of a boronic acid group
during a reaction involving other functional groups. Moreover, it would not complicate the
purification of the product as there would be no sensing molecule in the reaction mixture.
The natural anthraquinone alizarin, a weakly fluorescent compound in its free form,
becomes highly fluorescent when bound to a boronic acid via its 1,2-diol (Figure 1).2 This
property has been more commonly exploited with the water-soluble derivative alizarin red
S (ARS) (Figure 1), and the mechanism of the involved reaction has been studied in detail.3
In particular, ARS has been used for the direct visualization of a boronic acid group on a
cellulose surface4 and on glass slides5 by confocal microscopy.
During our synthetic work involving boronic acids (prior to the work described in chapter
3), we were interested to see whether the commercially available and inexpensive alizarin
(not to be confused with alizarin Red S) would be of help in visualizing boronic acids on
TLC plates. Alizarin has been used in combination with concentrated H2SO4 for the
fluorescent detection of boric acid on spot plates.6 Alizarin is also known as a TLC staining
reagent for the detection of cations (e.g., Al(III)), which requires an after-treatment with
ammonia.7 In this chapter, we introduce alizarin as a TLC staining reagent for the detection
of boronic acids.
Sensitive TLC detection of boronic acids using alizarin
61
Figure 1. Complexation of alizarin and alizarin Red S with boronic acids.
2 Results and discussion
2.1 Method development
Preliminary tests looked promising: after brief immersion of a TLC plate in a saturated
methanolic alizarin solution and spontaneous evaporation of the solvent, spots that
contained phenylboronic acid (PBA) turned yellow, whereas the rest of the plate turned
pink. Under 365 nm light, the spots showed an intense yellow fluorescence on a dark
background, even though it is mentioned in the article of Barder and Buchwald that alizarin
would require a significantly higher excitation wavelength (~550 nm).1
The detection of 2 µL spots of PBA solutions was optimized. A 50 mM solution of Na2CO3
in 50% aqueous MeOH has previously been used for measuring the affinity of alizarin for
boronic acids attached to solid supports.8 Solubility tests, moreover, showed that 1 mM was
near the saturation concentration of alizarin in MeOH. The following solutions, therefore,
were tested for the detection of PBA spots on a TLC plate: 1 mM alizarin in 1) 20 mM
Na2CO3 in water/MeOH 20:80, 2) MeOH, 3) 20 mM HCl in water/MeOH 20:80. The most
intense fluorescence was observed when pure MeOH was used (Figure 2A). Ethanol,
Alizarin: X = H-
O
O
X
OBOR
O
O
X
OB-O ROH
O
O
X
OH
OHOHBHOR
-
OHB-HO ROH
+
+ H2O - H++ H2O - H+- H2O
Alizarin Red S: X = SO3Na
Chapter 4
62
EtOAc, DCM, THF and acetone were also tested, and all solvents gave similar results in
terms of sensitivity: 2 µL spots of PBA solutions at concentrations down to 0.1 mM (24 ng)
could be easily detected (Figure 2A). Acetone was selected as the preferred solvent as it is
inexpensive, relatively safe, and the solubility of alizarin was higher in acetone than in the
other solvents (except THF). The detection of 2 µL spots of 1 mM PBA was subsequently
tested with alizarin solutions at different concentrations in acetone: 10 mM, 1 mM, 0.1 mM,
and 0.01 mM. The most intense fluorescence was observed with the initial concentration of
1 mM (Figure 2B). Staining the TLC plate using a piece of cotton, or dipping it in the
alizarin solution for 1 s, 5 s or 10 s resulted in the same sensitivity. For unknown reasons,
spraying the TLC plate with the alizarin solution did not result in any color change of the
TLC plate.
The following method was therefore chosen for subsequent experiments:
- Briefly dip the TLC plate in a 1 mM alizarin solution in acetone.
- Let dry and wait until the TLC plate is pink.
- Observe under 365 nm light.
After PBA was spotted in triplicate at concentrations ranging from 0.01 mM to 100 mM,
staining using the chosen method indicated that the limit of detection of PBA was between
0.01 mM and 0.1 mM. Measurement of the mean spot brightness (using graphics editor
GIMP) suggested that the developed method even enabled semi-quantification of PBA in
the spotted solutions (Figure 2C).
Sensitive TLC detection of boronic acids using alizarin
63
Figure 2. Development and semi-quantitative aspect of the method. Selections from original photographs of TLC
plates spotted with PBA (2 µL, various concentrations), briefly immersed in alizarin solution (various
compositions), left to dry and placed under 365 nm light.
10 1 0.1 0.01Alizarin conc. (mM)
C. Semi-quantitative aspect
A. Effect of acid, base and solvent used for the alizarin solution
B. Effect of the alizarin concentration in acetone
1
0.1
0.01PBA conc. (mM)
1 mM alizarin in:
10 1 0.1 0.01100PBA conc. (mM)
0
50
100
150
200
250
0.001 0.01 0.1 1 10 100
Mea
n sp
ot b
right
ness
PBA concentration (mM)
0
EtOH EtOAc DCM THF AcetoneMeOH
HClMeOHNa2CO3 MeOH
Chapter 4
64
2.2 Scope and selectivity
The ability of this method to detect diverse boronic acids and derivatives was tested (Figure
3A), and all compounds appeared yellow-orange. Electron-deficient boronic acids produced
a slightly more intense color than electron-rich boronic acids, which is probably due to the
pKa dependence of the boronic acid-diol complexation.9 PBA pinacol ester was revealed
with the same fluorescence intensity as PBA, which can be explained by the much higher
affinity of PBA for alizarin than for aliphatic diols,10 resulting in displacement of pinacol
by alizarin. Trifluoroborates were revealed with the same fluorescence intensity as their
corresponding boronic acids. PBA N-methyliminodiacetate (MIDA) ester could also be
detected, although it gave a much weaker fluorescence than PBA itself. The flavonoid
reagent diphenylborinic acid 2-aminoethyl ester (DPBA) gave a distinct orange
fluorescence upon contact with alizarin. Under neutral conditions the catechol group of
alizarin, like the catechol group of some flavonoids (e.g., rutin), probably gives a negatively
charged tetrahedral complex with DPBA after loss of ethanolamine.11 The difference in
fluorescence suggests that alizarin is able to distinguish between boronic and borinic esters.
The selectivity of this method towards boronic acids was tested by spotting compounds
with various functional groups on a TLC plate (Figure 3B). The fluorescence from these
compounds was very weak compared to the fluorescence from boronic acids at a 10-fold
lower concentration, and none of the tested boron-free compounds resulted in the
characteristic yellow spots observed in presence of the tested boron-containing compounds.
Sensitive TLC detection of boronic acids using alizarin
65
Figure 3. Scope and selectivity of the developed staining reagent. Selections from original photographs of TLC
plates spotted with diverse boronic acids and derivatives (10 mM in MeOH*, 2 µL; *MIDA ester: dry THF) (A)
and diverse boron-free compounds (100 mM in DCM, 2 µL) (B), briefly immersed in 1 mM alizarin solution in
acetone, left to dry and placed under 365 nm light.
BHO OH
BHO OH
CF3
BHO OH
COOH
BHO OH
OCH3
BOHHO
BOHHO
BOO
NH2
NH2
OB
PhPh
B
N
OOO OB F
F
FB F
F
F
K+ K+
CO2H OH PhO
NO2N3
CyN
CN
Cy
SH
Br
6
6 6
6
10
96
7 OH
Ph NH2
Ph CO2H
Ph CN
Tol SO3H
Ph BrPh Cl Ph I
PPh3
I5
Al(NO3)3.9H2O
A. Boronic acids and derivatives: 2 μL, 10 mM
B. Boron-free compounds: 2 μL, 100 mM
Chapter 4
66
2.3 Applications
The applicability of this method to the analysis of reactions involving boronic acids was
tested in 3 examples (Figure 4).
Example 1. Monitoring the consumption of a boronic acid during a classical Suzuki reaction. The reaction between 4-(trifluoromethyl)phenylboronic acid and bromobenzene
was carried out, based on the procedure described by Yang et al.12 A sample was spotted
every hour on a TLC plate. TLC analysis using 254 nm light did not give any indication
about the progress of the reaction. TLC analysis using the proposed method, however,
showed a clear decrease of the concentration of boronic acid over time. After 5 h, the
boronic acid was not detected anymore, which suggested that full conversion had been
obtained.
Example 2. Checking the completion of a ligand-free Suzuki reaction. The reaction
between PBA and 1-bromonaphthalene was carried out, using the modified procedure of
Liu et al.13 After 4 h of reaction, the reaction mixture was partitioned between DCM and
brine, and both layers were analyzed by TLC. TLC analysis using 254 nm light was
inconclusive. TLC analysis using the proposed method, however, showed that neither the
DCM layer nor the brine layer contained any significant amount of PBA, and that full
conversion had been obtained.
Example 3. Visualizing the transformation of a boronic acid into another boronic acid. The esterification of 4-carboxyphenylboronic acid in acidic ethanol was carried out, using
the procedure of Davison et al.14 After 1 h of reaction, the reaction mixture was analyzed by
TLC. TLC analysis using 254 nm light showed only weak spots that corresponded to
starting material and a new compound. TLC analysis using the proposed method, however,
showed the same spots with an intense yellow fluorescence. Apart from a much easier
observation, this also directly indicated that the new compound contained a boronic acid
moiety.
An additional example is shown in chapter 3, page 45 (without UV).
Sensitive TLC detection of boronic acids using alizarin
67
Figure 4. Applications of the developed staining reagent. Selections from original photographs of TLC plates
spotted with the indicated solutions, briefly immersed in 1 mM alizarin solution in acetone, left to dry and placed
under 365 nm light.
Br
BHO OH
Pd(PPh3)4Na2CO3
DMEWaterRefluxunder Ar
254 nm 365 nm
1) Dippedin alizarinsolution
2) Allowedto dry
CF3
CF3
1 2
SpotsA. AB. Boric acid0, 1, 2, 3, 5. h of reaction
3 50A B
A
1 2 3 50A B
BrB
HO OH
Pd(OAc)2Na2CO3
AcetoneWaterRT, air2
1
1 2
Spots1. 12. 2 3 4 5 1 2 3 4 5
254 nm 365 nm
1) Dippedin alizarinsolution
2) Allowedto dry
EthanolConcd HCl
Reflux
1 2
Spots1. Starting material2. Reaction mixture
254 nm 365 nm
BOHHO
OHO
BOHHO
OO
1 2
1) Dippedin alizarinsolution
2) Allowedto dry
EluentHeptane/EtOAc7:3
EluentHeptane/EtOAc2:3
EluentEtOAc
3. DCM layer4. Brine layer5. Boric acid
Example 1
Example 2
Example 3
Chapter 4
68
3 Experimental section
Example 1. A mixture of 4-(trifluoromethyl)phenylboronic acid (190 mg, 1 mmol),
bromobenzene (0.15 mL, 1.4 mmol) and Na2CO3 (318 mg, 3 mmol) in dimethoxyethane (6 mL) and water (1.6 mL) was degassed with Argon for 10 min and heated to reflux.
Pd(PPh3)4 (58 mg, 0.05 mmol) was added and the reaction mixture was stirred under reflux.
Immediately after adding the catalyst, and then after 1 h, 2 h, 3 h and 5 h, aliquots (~ 0.2 mL, except ~ 0.5 mL after 5 h) of the reaction mixture were partitioned between
Et2O (1 mL) and 0.1 M HCl (2 mL), and the Et2O layer was spotted on a TLC plate.
Example 2. A mixture of phenylboronic acid (122 mg, 1 mmol), 1-bromonaphthalene (0.15 mL, 1.1 mmol), Na2CO3 (212 mg, 2 mmol) and Pd(OAc)2 (~ 1 mg) in acetone (3 mL)
and water (3.5 mL) was stirred at RT in air.
Example 3. 4-Carboxyphenylboronic acid (92 mg, 0.55 mmol) was dissolved in absolute
ethanol (2.5 mL). A mixture of absolute ethanol (2.5 mL) and concentrated HCl (2 drops)
was added, and the solution was heated to reflux.
4 Conclusion
The difficult monitoring of boronic acids during organic reactions can be simply overcome
by briefly dipping TLC plates in a 1 mM alizarin solution, to reveal boronic acids on TLC
plates in a selective and sensitive manner.
Sensitive TLC detection of boronic acids using alizarin
69
5 References
1. T. E. Barder and S. L. Buchwald, Benchtop monitoring of reaction progress via visual recognition with
a handheld UV lamp: In situ monitoring of boronic acids in the Suzuki-Miyaura reaction, Organic
Letters, 2007, 9, 137-139.
2. Y. Kubo, A. Kobayashi, T. Ishida, Y. Misawa and T. D. James, Detection of anions using a fluorescent
alizarin-phenylboronic acid ensemble, Chemical Communications, 2005, 2846-2848.
3. J. W. Tomsho and S. J. Benkovic, Elucidation of the mechanism of the reaction between phenylboronic
acid and a model diol, alizarin red S, Journal of Organic Chemistry, 2012, 77, 2098-2106.
4. D. Zhang, K. L. Thompson, R. Pelton and S. P. Armes, Controlling deposition and release of polyol-
stabilized latex on boronic acid-derivatized cellulose, Langmuir, 2010, 26, 17237-17241.
5. A. E. Ivanov, A. Kumar, S. Nilsang, M. R. Aguilar, L. I. Mikhalovska, I. N. Savina, L. Nilsson, I. G.
Scheblykin, M. V. Kuzimenkova and I. Y. Galaev, Evaluation of boronate-containing polymer brushes
and gels as substrates for carbohydrate-mediated adhesion and cultivation of animal cells, Colloids
and Surfaces B: Biointerfaces, 2010, 75, 510-519.
6. L. Szebellédy and S. Tanay, Beiträge zum Nachweis der Borsäure mit Alizarin, Zeitschrift für
Analytische Chemie, 1936, 107, 26-30.
7. E. Stahl, Thin-layer chromatography: a laboratory handbook, Springer Berlin Heidelberg, 1969.
8. P. J. Duggan and D. A. Offermann, The preparation of solid-supported peptide boronic acids derived
from 4-borono-L-phenylalanine and their affinity for alizarin, Australian Journal of Chemistry, 2007,
60, 829-834.
9. J. A. Peters, Interactions between boric acid derivatives and saccharides in aqueous media: Structures
and stabilities of resulting esters, Coordination Chemistry Reviews, 2014, 268, 1-22.
10. G. Springsteen and B. Wang, A detailed examination of boronic acid-diol complexation, Tetrahedron,
2002, 58, 5291-5300.
11. P. Matteini, G. Agati, P. Pinelli and A. Goti, Modes of complexation of rutin with the flavonoid reagent
diphenylborinic acid 2-aminoethyl ester, Monatshefte fur Chemie, 2011, 142, 885-893.
12. C. Yang, R. Edsall Jr, H. A. Harris, X. Zhang, E. S. Manas and R. E. Mewshaw, ERβ Ligands. Part 2:
Synthesis and structure–activity relationships of a series of 4-hydroxy-biphenyl-carbaldehyde oxime
derivatives, Bioorganic & Medicinal Chemistry, 2004, 12, 2553-2570.
13. L. Liu, Y. Zhang and B. Xin, Synthesis of biaryls and polyaryls by ligand-free Suzuki reaction in
aqueous phase, Journal of Organic Chemistry, 2006, 71, 3994-3997.
14. J. Z. Davison, W. D. Jones, H. Zarrinmayeh and D. M. Zimmerman, United States Pat.,
US2003/225266 A1, 2003.
Chapter 4
70
71
Chapter 5Selective on-line detection of boronic acids and derivatives in HPLC eluates by post-column reaction with alizarin
An HPLC method for the rapid and selective detection of boronic acids in complex
mixtures was developed. After optimization experiments at an HPLC flow rate of 0.40
mL/min, the HPLC-separated analytes were mixed post-column with a solution of 75 μM
alizarin and 0.1% triethylamine in ACN, which was delivered at a flow rate of 0.60
mL/min. The reaction between alizarin and boronic acids occurred in a reaction coil of
dimensions of 3.5 m × 0.25 mm at a temperature of 50 °C, resulting in fluorescent
complexes that were detected as positive peaks by a fluorescence detector (exc 469 nm and
em 610 nm). The method enabled the selective detection of various boronic acids and
derivatives, with a limit of detection of phenylboronic acid of 1.2 ng or 1 μM. It could
successfully monitor the progress of two organic reactions involving boronic acid-
containing compounds, and provided useful insights into the course of the reactions.
This chapter is a slightly modified version of the following manuscript: F. Duval, P.A. Wardani, H. Zuilhof, T.A. van Beek, Selective on-line detection of boronic acids and derivatives in high-performance liquid chromatography eluates by post-column reaction with alizarin, Journal of Chromatography A, revision submitted.
10
20
30
40
4 6 8 10
mA
U
min
Absorbance at 254 nm
Fluorescence at 610 nm, with alizarinFluorescence at 610 nm, without alizarin
Chapter 5
72
1 Introduction
To facilitate the analysis of reaction mixtures that contain boronic acids, we developed a
method to stain boronic acids on silica TLC plates in a selective and sensitive way, using
the natural dye alizarin (chapter 4). Alizarin shows little fluorescence by itself, but when it
interacts with a boronic acid in suitable conditions, a strongly fluorescent boronic ester is
formed (chapter 4, Figure 1). When the alizarin-stained and dried TLC plate is placed under
a UV lamp (365 nm), yellow fluorescent spots reveal the presence of boronic acids.
This method is limited, however, to TLC analysis on silica plates. TLC is a useful tool and
the first line of approach in organic synthesis. However, its limited separation capacity
renders it sometimes essential to analyze a complex reaction mixture by reversed-phase
high-performance liquid chromatography (RP-HPLC). For the preparative separation of
boronic acids, moreover, preparative RP-HPLC is superior to preparative TLC.
On-line HPLC post-column derivatization has received much attention, as it is a useful tool
for the detection of specific analytes in complex mixtures.1 In this approach an HPLC
eluate and reagent solution are mixed together, and the resulting conversion of the analyte
of interest allows its selective detection. Our group has developed different on-line HPLC
detection methods for radical scavenging compounds.2 Such compounds either react post-
column with a stable free radical such as DPPH or ABTS+,3 or prevent the oxidation of a
sensitive probe4 resulting in a change in absorbance. Such methods proved versatile and
could be used preparatively5 and applied to complex mixtures.6
Our experience with TLC detection of boronic acids using alizarin and on-line HPLC
detection based on post-column reactions led to the idea of combining both techniques with
the aim of selectively detecting boronic acids in HPLC eluates to facilitate analysis or
subsequent preparative HPLC of reaction mixtures containing boronic acids. In this chapter,
the development and application of such on-line HPLC detection methodology for boronic
acids are presented.
Selective on-line detection of boronic acids in HPLC eluates
73
2 Experimental
2.1 Chemicals and their abbreviations
For the analytical experiments, the following chemicals were used: alizarin (97%) and
triethylamine (TEA) (99%) from Acros Organics; acetonitrile (ACN) (HPLC-S) from
Biosolve BV (Valkenswaard, The Netherlands); methanol (MeOH) (CHROMASOLV®,
for HPLC, ≥99.9%), acetic acid (ACS Reagent ≥99.7%), phosphate-buffered saline pH 7.4
(PBS) (cat. no P3813), citric acid, 2-(cyclohexylamino)ethanesulfonic acid (CHES), 4-(2-
hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES), sodium hydroxide, alizarin Red S
(ARS) (certified by Biological Stain Commission), phenylboronic acid (PBA) (purum,
≥97.0%), phenylboronic acid pinacol ester, phenylboronic acid N-methyliminodiacetic acid
(MIDA) ester (97%), potassium phenyltrifluoroborate (95%), 4-trifluoromethyl-
phenylboronic acid (≥95.0%), 4-carboxyphenylboronic acid, 4-methoxyphenylboronic acid
(≥95.0%), 3-aminophenylboronic acid monohydrate (98%), butylboronic acid (97%),
potassium butyltrifluoroborate (95%), benzoic acid (ACS reagent, ≥99.5%), nitrobenzene
(ACS reagent, ≥99.0%), acetophenone (puriss. p.a., ≥99.0%), 4-hydroxybenzamide (98%),
chlorobenzene (puriss. p.a., ACS reagent, ≥99.5%), bromobenzene ( ≥99.5%), p-
toluenesulfonic acid monohydrate (ACS reagent, ≥98.5%), aniline (ACS reagent, ≥99.5%),
phenol (unstabilized, ReagentPlus®, ≥99%) and 4-methoxybenzonitrile (99%) from Sigma-
Aldrich; boric acid from an unknown source, sodium borate made in situ from boric acid
and sodium hydroxide; benzaldehyde (≥98%) from Fisher Scientific and 4-
fluorobenzaldehyde (98%) from VWR International. Ultrapure water (= “water”) was
produced using a Milli-Q Integral 3 system from Millipore (Molsheim, France). For the
esterification reaction, the following chemicals were used: 4-carboxyphenylboronic acid
(CPBA) from Sigma-Aldrich, absolute ethanol (AnalaR NORMAPUR® ACS) from VWR
International and hydrochloric acid (37%) from Acros Organics. For the thiol-ene reaction,
the following chemicals were used: (4-allylaminocarbonyl)phenylboronic acid (AACPBA)
(97%) and N-α-t-butyloxycarbonyl-L-cysteine (Boc-Cys) (99%) from ABCR (Karlsruhe,
Germany), 2,2-dimethoxy-2-phenylacetophenone (DMPA) (99%) from Sigma-Aldrich,
HPLC-grade tetrahydrofuran (unstabilized, HPLC-S) from Biosolve BV.
Chapter 5
74
2.2 Off-line fluorescence spectrometry
Solutions containing alizarin (or ARS) and/or boronic acids were placed in quartz cuvettes
and analyzed with an Edinburgh Instruments FLS900 Fluorescence spectrometer. A Xe900
lamp was used for excitation and an R928 photomultiplier tube was used for detection.
Excitation spectra were measured with λem 610 nm and emission spectra were measured
with λex 469 nm. Measurements were performed about 15 minutes after mixing alizarin (or
ARS) with boronic acid solutions. Detailed composition of these solutions can be found in
the corresponding figure captions.
2.3 On-line HPLC instrumental set-up
Figure 1. Instrumental set-up for HPLC on-line detection of boronic acids using alizarin.
The instrumental set-up is depicted in Figure 1. The HPLC delivery system consisted of a
WellChrom HPLC Pump K-1001 and a WellChrom Solvent Organizer K-1500 (Knauer,
Germany). The manual injector was a Multiport Streamswitch MUST HP 6 (Spark Holland,
Emmen, the Netherlands) with a loop of 10 μL. Separations were carried out on an Alltima
C18 HPLC column (5 µm, 150 mm × 3 mm i.d., pore size 100 Å, Alltech Associates Inc.,
Deerfield, IL), and the HPLC flow rate was set at 0.40 mL/min. UV detection was carried
out using a 2487 Dual λ Absorbance Detector (Waters, Milford, MA) set at a detection
wavelength of 254 nm. Post-column delivery of the alizarin solution was achieved as
follows: the alizarin solution was displaced from a 150 mL Superloop™ (GE Healthcare
Bio-Sciences AB, Uppsala, Sweden) by infusing water into the opposite end of the
HPLC delivery system
Fluorescence detector
Alltima C18 column
T-pieceUV
detectorAlizarinin Superloop
Recorder
Recorder
HPLCpump
Waste
Solvent reservoirs
Injector
Rea
ctio
n co
il in
wat
er b
ath
Selective on-line detection of boronic acids in HPLC eluates
75
Superloop™ by means of a Gynkotek High Precision Pump Model 300. The flow rate of
this HPLC pump is further described as “alizarin flow rate”. The tested reaction coils were
made of PEEK tubing (VICI AG International, Schenkon, Switzerland). Post-column
fluorescence detection was done using an FP-1520 Intelligent Fluorescence Detector
(JASCO, Tokyo, Japan), λex 469 nm and λem 610 nm. Components of the set-up were
connected with tubing of various materials and dimensions (details in appendix, Table 1).
2.4 Flow injection analysis: effect of coil temperature and length
2.4.1 Instrumental set-up
Flow injection analysis experiments were carried out with the same set-up as described in
paragraph 2.3 and depicted in Figure 1 with one difference: the HPLC column was mounted
before the injector, as it was used only for keeping the system at the desired pressure for
proper functioning of the HPLC pump.
2.4.2 Operating conditions
The carrier flow consisted of ACN/water 1:1 (v/v) at a flow rate of 0.40 mL/min. The
Superloop™ was filled with a solution of 3.13 µM alizarin and 0.10% TEA in ACN. The
alizarin flow rate was set at 0.80 mL/min. 313 µM PBA in MeOH was injected at each new
temperature or coil length.
2.4.3 Effect of coil temperature
The temperature of the reaction coil was changed by immersing the latter in a water bath on
a heating plate equipped with a thermostat. Four different temperatures were tested for the
detection of PBA: 22 °C (RT), 30 °C, 40 °C and 50 °C. The coil had dimensions of 3.50 m
× 0.25 mm. All other conditions described in 2.4.2 were fixed.
Chapter 5
76
2.4.4 Effect of coil length
Reaction coils of identical internal diameters (0.25 mm) and three different lengths (3.5, 6.0
and 10.0 m) were tested for the detection of PBA. All other conditions described in 2.4.2
were fixed.
2.5 On-line HPLC
2.5.1 Instrumental set-up
All experiments in this section were carried out using the set-up and conditions as described
in 2.3 and depicted in Figure 1.
2.5.2 Effect of alizarin concentration on S/N ratio
The HPLC mobile phase consisted of ACN/water 2:3 (v/v). The Superloop™ was filled
with a solution of alizarin and 0.10% TEA in ACN. The alizarin concentration ranged from
1 to 300 µM, and the alizarin flow rate was set at 0.60 mL/min. At each alizarin
concentration, PBA solutions at concentrations ranging from 1 to 300 µM in MeOH were
injected, in triplicate (unless stated otherwise). The S/N ratio was determined as follows:
the signal S was defined as the difference between peak height and average signal height of
the baseline, and the noise N was defined as 3× standard deviation of the baseline around
the elution time of the analyte of interest for a blank injection (solvent used for dissolving
the analyte).
2.5.3 Effect of alizarin flow rate on S/N ratio
The HPLC mobile phase consisted of ACN/water 2:3 (v/v). The Superloop™ was filled
with a solution of 75 µM alizarin and 0.10% TEA in ACN. The alizarin flow rate was
varied from 0.0 to 1.0 mL/min with increments of 0.10 mL/min. At each alizarin flow rate,
300 µM PBA in MeOH was injected, in triplicate. The S/N ratio was determined as in
2.5.2.
Selective on-line detection of boronic acids in HPLC eluates
77
2.5.4 Proof-of-principle, scope and selectivity
The HPLC mobile phase consisted of a gradient of 20 to 50% ACN in water. The
Superloop™ was filled with a solution of 75 µM alizarin and 0.10% TEA in ACN, and the
alizarin flow rate was set to 0.60 mL/min. Solutions of various compounds in ACN/water
1:5 (see structures and concentrations in Figure 5) were injected in identical conditions
(except compound concentration when a change was deemed necessary for sensitivity
reasons). When it was necessary to check whether the compounds were autofluorescent, the
same experiments were repeated without any alizarin flow. The S/N ratio was determined
as in 2.5.2.
2.5.5 Estimation of the LOD
The limit of detection for PBA was calculated in two ways: (1) by finding the concentration
in a calibration curve using peak heights giving a signal of 3× the SD of the average
baseline of a blank injection at the elution time of PBA; (2) by injecting samples of 0.3, 1.0
and 3.0 µM of PBA and determining which concentration still gives a visible peak.
2.5.6 Applications
Acid-catalyzed esterification of CPBA CPBA (92 mg, 0.55 mmol) was dissolved in absolute ethanol (2.5 mL). HCl 37% (2 drops)
was diluted in absolute ethanol (2.5 mL) and added to the previous solution. The obtained
solution was stirred and heated to reflux.
The HPLC mobile phase consisted of MeOH/water 2:3 (v/v) with 0.1% acetic acid. The
Superloop™ was filled with a solution of 75 µM alizarin and 1% TEA in ACN, and the
alizarin flow rate was set to 0.60 mL/min. Samples were taken from the reaction mixture
just before reflux started, and 0.5, 1, 2, 3, 6 and 21 h after reflux started, diluted 100 times
in MeOH/water 2:3 (v/v) and injected.
Chapter 5
78
Thiol-ene reaction between AACPBA and Boc-Cys AACPBA (125 mg, 0.61 mmol, 1 eq), Boc-Cys (675 mg, 3.05 mmol, 5 eq) and DMPA
(156 mg, 0.61 mmol, 1 eq) were dissolved in THF (3 mL). The obtained solution was
stirred under 365 nm irradiation (using a portable UV lamp as used for viewing of TLC
plates, Camag, Switzerland).
The HPLC mobile phase consisted of a MeOH/water 1:1 (v/v) with 0.1% acetic acid. The
Superloop™ was filled with a solution of 75 µM alizarin and 1% TEA in ACN, and the
alizarin flow rate was set to 0.60 mL/min. Samples were taken from the reaction mixture
before irradiation and 1, 2, 5, 10, 20, 40, 80 and 160 min after irradiation was started,
diluted 100 times in MeOH/water 1:1 (v/v) and injected.
3 Results and discussion
3.1 Off-line fluorescence spectrometry
3.1.1 Choice of dye and solvent
The natural dye alizarin has been used for the TLC detection of boronic acids as described
in chapter 4. However, the water-soluble derivative alizarin red S (ARS) is more widely
used for its binding with boronic acids. To test the fluorescence detection of boronic acids
in solution, preliminary experiments were therefore performed with ARS. Excitation and
emission scans of ARS only and ARS/PBA mixtures (appendix, Figure 1) showed that 1)
ARS had a significant autofluorescence, 2) the most efficient excitation wavelength for
PBA detection was 469 nm (which was used in all subsequent experiments), and 3)
subtraction of the ARS emission spectrum from the ARS/PBA emission spectrum resulted
in a spectrum with a fluorescence intensity maximum at 610 nm (which was used in
subsequent experiments that required the choice of an emission wavelength). Next,
mixtures of ARS with various boronic acids were analyzed in the same way, and showed
high variations in fluorescence intensity (appendix, Figure 2). The electron-deficient
boronic acids resulted in high fluorescence intensities (e.g., 4-trifluoromethylphenylboronic
Selective on-line detection of boronic acids in HPLC eluates
79
acid), and the electron-rich boronic acids resulted in low fluorescence intensities (e.g., 4-
methoxyphenylboronic acid, MPBA). Based on these results, the immediate goal was to
improve the detection of electron-rich boronic acids, and MPBA was taken as a model
compound. Because we had positive experiences with alizarin, alizarin was compared with
ARS for the detection of MPBA: ARS, alizarin, ARS/MPBA and alizarin/MPBA solutions
were analyzed. The obtained fluorescence spectra (appendix, Figure 3) revealed that the
detection of MPBA was improved when using alizarin. Alizarin, moreover, had a much
lower autofluorescence than ARS. As this constituted a significant advantage in the on-line
HPLC detection of boronic acids due to the much higher relative change in fluorescence
intensity, alizarin was used in all further experiments.
Additional off-line experiments with alizarin showed that an increase of the organic solvent
percentage resulted in an increase of the fluorescence intensity (appendix, Figure 4): the
less water, the higher the intensity. Because HPLC eluates typically contain water, the final
percentage of water after mixing eluate and alizarin solution was minimized by dissolving
alizarin in 100% organic solvent. MeOH and ACN were both considered but ACN was preferred because 1) ACN is less viscous than MeOH, hence it causes a lower
backpressure than MeOH does, and 2) MeOH is prone to give bubble formation upon
mixing with a water-containing eluate, which might disturb the fluorescence detection after
the low-pressure coil. The same would be true if water was to be mixed with a MeOH-
containing eluate; this is an additional advantage of using alizarin instead of ARS, as ARS
would have to be dissolved in water.
Chapter 5
80
3.1.2 Effect of pH
To improve the composition of the alizarin solution, the effect of pH on the fluorescence
intensity of alizarin/PBA mixtures was studied (a similar experiment has been published,
where ARS was used instead of alizarin7), using an extended-range citrate/HEPES/CHES
buffer system.8 The largest difference in fluorescence intensity (= the highest sensitivity)
between alizarin only and alizarin/PBA was obtained at pH values between 7 and 8
(appendix, Figure 5). Because ACN was chosen as solvent for the alizarin solution, it was
not possible to make a buffer at a specific pH. Instead, the effect of triethylamine (TEA) on
the fluorescence intensity of alizarin/PBA mixtures was explored. Conditions were chosen
in such a way that they would roughly mimic on-line HPLC, and the analyzed solutions
consisted of two parts of identical volumes: 1) “alizarin solution mimic”: 20 μM alizarin
solution in ACN that contained 0%, 0.1%, 1% or 10% TEA, and 2) “HPLC eluate mimic”:
200 μM PBA in 50% water and 50% ACN or MeOH. Independent of using MeOH or ACN
in the “HPLC eluate mimic”, the highest sensitivity was achieved when the “alizarin
solution mimic” contained 0.1% TEA (appendix, Figure 6).
The addition of small amounts of acid, e.g., acetic acid, to the mobile phase is often
necessary for obtaining a good retention and separation of weakly acidic analytes. Thus, the
same experiment as above was performed, with one difference: the “HPLC eluate mimic”
was made acidic via the use of 0.1% acetic acid. When the “alizarin solution mimic”
contained 0% or 0.1% TEA, the resulting fluorescence intensity was much weaker than in
the previous experiment. However, when the “alizarin solution mimic” contained 1% TEA,
the resulting fluorescence intensity could be brought back to the maximum observed in the
previous experiment (appendix, Figure 7).
Based on the results of the off-line experiments, the following choices were made: the post-derivatization solution would consist of alizarin in ACN, to which 0.1% TEA (in case of a neutral eluate) or 1% TEA (in case of an eluate that contains small percentages of acidic modifiers) would be added.
Selective on-line detection of boronic acids in HPLC eluates
81
3.2 Flow injection analysis: effect of coil temperature and length
Next, the efficacy of the above solution to enable on-line detection of boronic acids was
tested by flow injection analysis. To ensure a proper operation (pulse dampening) of the
HPLC pump providing the carrier flow, an HPLC column was mounted between the pump
and the injector. Columns with diameters of 2.1 mm, 3.0 mm and 4.6 mm were available,
with respective recommended flow rates of 0.20 mL/min, 0.40 mL/min and 1.0 mL/min. A
flow rate of 0.40 mL/min was preferred, as the lowest flow rate may result in a poor
detection sensitivity and significant peak broadening, and the highest flow rate would result
in a too high pressure for the Superloop™ (maximum 10 bar) and a high consumption of
HPLC solvents and alizarin solution. A column of 3.0 mm diameter at a flow rate of 0.40 mL/min was therefore chosen.
The sensitivity of the method depends on the conversion of alizarin and the specific boronic
acid into a fluorescent alizarin-boronic acid complex. At given concentrations of alizarin
and boronic acid, two major factors influencing this conversion are temperature and
reaction time. First, the detection of PBA was tested with a coil of dimensions 3.5 m × 0.25
mm at four temperatures (22 °C to 50 °C). This showed that the peak height of PBA
increased linearly with temperature (Appendix, Figure 8). A coil temperature of 50 °C was therefore used in all further experiments. A higher temperature was not considered
because of problems associated with post-coil ACN volatility. Second, 0.25 mm i.d. coils of
three different lengths (3.5, 6.0 and 10.0 m) were compared. Coil dimensions directly
define the residence time, hence the time that alizarin and boronic acid have for reacting
with each other. Upon increasing coil length, at 50 °C only a minor increase of PBA peak
height was observed: for coil lengths of 3.5, 6.0 and 10.0 m (resulting in respective
residence times of about 9, 15 and 25 s), respective peak heights were 8.3, 8.8 and 8.9 milli
arbitrary units (mAU). Thus, to avoid (too) high backpressures in the Superloop™ as well
as peak broadening, coil dimensions of 3.5 m × 0.25 mm were used in all subsequent
experiments.
Chapter 5
82
3.3 On-line HPLC
3.3.1 Effect of alizarin concentration on S/N ratio
The on-line detection of PBA is based on the S/N ratio of the produced fluorescent peak,
which is directly affected by the alizarin concentration of the solution that mixes with the
HPLC eluate. A too low concentration would result in a poor signal, and a too high
concentration would result in too much noise. In order to find the optimum alizarin
concentration, the S/N ratio for the detection of PBA (at concentrations from 1 to 300 μM)
was determined at fixed flow rates: 0.40 mL/min for the HPLC mobile phase and 0.60
mL/min for the alizarin solution. 0.60 mL/min was selected as this flow was expected to be
near the optimum on the basis of peak broadening, mixing, sensitivity and back pressure
issues. The alizarin concentration was varied from 1 to 300 μM. As a result, the highest S/N ratio was obtained with an alizarin concentration of 75 μM (Figure 2A; see
appendix, Figure 9 for lower PBA concentrations), which was used in all subsequent
experiments.
3.3.2 Effect of alizarin flow rate on S/N ratio
An increase of the alizarin flow rate can give rise to two counteracting effects: (1) an
increase of the alizarin concentration and of the percentage of organic solvent (ACN) when
the alizarin solution and HPLC eluate are mixed, hence a possible increase in fluorescence
intensity in presence of boronic acids; (2) a decrease of the boronic acid concentration and
of the amount of time that is available for alizarin and boronic acid to form a fluorescent
complex, hence a possible decrease in fluorescence intensity in presence of boronic acids.
In addition, there are peak broadening issues (higher flows are better) and mixing issues
(near-equal column and alizarin flows work best). In order to find the optimum alizarin
flow rate, the latter was varied from 0.0 to 1.0 mL/min with increments of 0.1 mL/min at
the optimized alizarin concentration of 75 µM. The S/N ratio was highest when an alizarin flow rate of 0.6 mL/min was used (Figure 2B), but flows of 0.5 or 0.7 mL/min are also
usable.
Selective on-line detection of boronic acids in HPLC eluates
83
Figure 2. A. Effect of alizarin concentration on S/N ratio. At each alizarin concentration, PBA was injected at
various concentrations and the signal-to-noise ratio (S/N) of each produced peak was determined. n = 3 except for
50 µM alizarin in which case n = 2. B. Effect of alizarin flow rate on S/N ratio. At each alizarin flow rate,
phenylboronic acid (300 μM, 10 μL) was injected in triplicate and the signal-to-noise ratio (S/N) of the produced
peak was determined.
1 3 10 30 50 75 100 300
(S/N
)/100
0
Alizarin concentration (μM)
1030100300
PBA conc.(μM)2
1
0
A
0
2
4
6
8
10
0.0 0.2 0.4 0.6 0.8 1.0
(S/N
)/100
0
Alizarin flow rate (mL/min)
B
Chapter 5
84
3.3.3 Proof-of-principle and semi-quantitative aspect of the developed method
Based on the results that have been presented so far, the following conditions were fixed:
- The post-column derivatization solution consisted of 75 μM alizarin and 0.1% TEAin ACN and was delivered at a flow rate of 0.60 mL/min (at an HPLC flow rate of 0.40mL/min).
- The reaction coil had dimensions of 3.5 m × 0.25 mm and was immersed in a waterbath at a temperature of 50 °C.
- The fluorescence detector was set at exc 469 nm, and em 610 nm.
Figure 3 shows the proof-of-principle chromatograms using the conditions described above.
PBA was injected, once with and once without post-column alizarin flow. The fluorescence
was measured in both cases, and PBA was detected by fluorescence only in presence of
alizarin. The on-line fluorescence detection resulted in an acceptable 35% peak broadening
as compared to the UV detection.
Figure 3. Proof-of-principle of the developed method. 300 μM PBA was injected, with (red) and without (black)
alizarin delivery after the absorbance detector (blue).
10
20
30
40
4 6 8 10
mA
U
min
Absorbance at 254 nm
Fluorescence at 610 nm, with alizarin Fluorescence at 610 nm, without alizarin
7.10 min
6.77 min
Selective on-line detection of boronic acids in HPLC eluates
85
The ability of the method to quantify PBA was evaluated using the chromatograms
(appendix, Figure 10) obtained from the optimization of the alizarin concentration (3.3.1),
when 75 μM alizarin was used. The peak area increased linearly with PBA concentration
(appendix, Figure 11). Based on a peak height calibration curve (Figure 4), the LOD (3×
SD of the noise) for PBA was estimated at 0.15 µM. When solutions of 0.3, 1 and 3 µM
of PBA were injected, the PBA peak corresponding with 1 µM was still visible, but not the
one of 0.3 µM. Thus conservatively, the LOD is set at 1 µM in combination with a 10 µL
loop. The minimal detectable amount of PBA is 1.2 ng, which gives the method ample
sensitivity for the analysis of boronic acids in synthetic mixtures.
Figure 4. Peak height calibration curve and estimation of the limit of detection (LOD) of PBA. Peak heights taken
from chromatograms in Figure S10 (+ replicates). n = 3 except for 3 µM PBA in which case n = 2.
y = 0.0741x
R² = 0.9997
0
0.2
0.4
0.6
0.8
0 2 4 6 8 10
Peak
hei
ght (
mA
U)
PBA concentration (µM)
3 × SDnoise = 0.011 LOD = 0.011/0.0741 Thus LOD = 0.15 µM
Chapter 5
86
3.3.4 Scope and selectivity
The final goal was the selective detection of all kinds of boronic acids and derivatives. To
evaluate the scope and selectivity of the developed method, various boronic acids, boronic
acid derivatives and boron-free compounds were injected. The S/N ratio observed for each
compound can be found in Figure 5. The S/N ratios of boronic acids and derivatives were
somewhat variable for the following reasons: (1) the complexation between boronic acid
and alizarin depends on the pKa of the boronic acid, which is affected by the presence of
electron-donating or electron-withdrawing groups;7 (2) the complexation between boronic
acid and alizarin depends on the pH,7 which can be affected by a basic or acidic group in
the boronic acid molecule; (3) in one single case (4-methoxyphenylboronic acid),
autofluorescence of the compound contributed to the signal (chromatograms of all analyzed
boronic acids in appendix, Figure 12). Regardless of the differences in fluorescence
intensities, all boronic acids and derivatives could be detected. They included compounds
that could not be detected by UV (such as butylboronic acid and boric acid), showing an
additional advantage of the developed method. In contrast, most boron-free compounds,
which were injected at much higher concentrations, were not or hardly detected. The boron-
free compounds that did give a signal (see Figure 5) were also injected in the absence of
alizarin, and proved to be autofluorescent (chromatograms of autofluorescent compounds
with and without alizarin in appendix, Figure 13). In other words, the false positives were
not due to complexation with alizarin, but to the intrinsic fluorescent properties of the
analyzed compounds.
Selective on-line detection of boronic acids in HPLC eluates
87
Figure 5. Scope and selectivity of the developed method. Various compounds were injected, and the signal-to-
noise ratio (S/N) of each produced peak was determined.
0
2
4
6
8
0
2
4
6
8
16
18
Boronic acids and derivatives
Other compounds
B
N
OOO O
BHO OH
OCH3
BOHHO
NH2
0.30 mM
0.30 mM
0.30 mM
0.30 mM
0.30 mM
0.30 mM
3.00 mM
3.00 mM
1.30 mM
1.20 mM
0.25 mM
1.20 mM
PhH
O
OH
NH2O3.00 mM
3.00 mM
3.00 mM
3.00 mM
3.00 mM
3.00 mM
3.00 mM
3.00 mM
3.00 mM
0.30 mM
0.30 mM
0.30 mM
(S/N
)/100
0(S
/N)/1
000
Chapter 5
88
3.3.5 Applications
The developed system was tested with real-life synthetic mixtures. The first example
concerned the monitoring of the acid-catalyzed esterification of CPBA (Figure 6). The
same experimental procedure as in chapter 4, in which this reaction was monitored by TLC,
was used. Samples were taken from the reaction mixture after various reaction times,
diluted and analyzed using the developed method. Figure 6 shows chromatograms at t = 0
and t = 21 h. At t = 0, only CPBA was detected, both by absorbance and fluorescence. At t
= 21 h, CPBA had almost completely disappeared and two major new absorbance peaks
appeared. The fluorescence chromatogram shows that only the second major peak (by
retention time) may correspond to a boronic acid compound, most probably the expected
product of the reaction. Figure 14 of the appendix shows the fluorescence peak heights of
starting material and product against reaction time.
The second example is the monitoring of the thiol-ene reaction between AACPBA and
Boc-Cys, as described in chapter 3, resulting in the formation of a new boronic acid-
containing molecule (Figure 7). Samples were taken from the reaction mixture after various
reaction times, diluted and analyzed using the developed method. Figure 7 shows parts of
chromatograms at selected reaction times. In the UV 254 nm chromatograms, many new
peaks appeared and it was hard to interpret the peak pattern. The new boronic acid-selective
detection, however, led to more insight: the newly formed UV 254 nm peak at 8.1 min
corresponded with a new fluorescence peak at 8.4 min in the post-column chromatogram
(0.3 min between UV and fluorescence detectors), strongly indicating that it was the
product of the reaction. The new UV 254 nm peak at 7.0 min, however, did not correspond
with any fluorescence peak: this indicated that a by-product (among others) was formed, in
which the boronic acid moiety had disappeared. Peak heights of starting material, product
and by-product were plotted against reaction time, and suggested that the formed product
(8.1 min in UV 254 nm chromatogram) degraded into the by-product (7.0 min in UV 254
nm chromatogram) (appendix, Figure 15).
Selective on-line detection of boronic acids in HPLC eluates
89
Figure 6. Monitoring of the acid-catalyzed esterification of CPBA. Samples were taken from the reaction
mixture at various reaction times, diluted and injected. Parts of chromatograms for selected reaction times.
16 18 20 22 24 26 28 min
16 18 20 22 24 26 28 min
0
20
40
60
80
0 2 4 6 8
mA
U
0
20
40
60
80
0 2 4 6 8
mA
U
Absorbance at 254 nm
Fluorescence at 610 nm
t = 0
t = 21 h
Boronic acid-containing
starting material (CPBA)
Boron-freeby-product Boronic acid-containing
product
Chapter 5
90
Figure 7. Monitoring of the thiol-ene reaction between AACPBA and Boc-Cys. Samples were taken from
the reaction mixture at various reaction times, diluted and injected. Parts of chromatograms for selected
reaction times.
0
50
100
150
200
0 2 4 6 8 10
mA
U
min0
50
100
150
200
0 2 4 6 8 10
mA
U
min
0
50
100
150
200
0 2 4 6 8 10
mA
U
min
t = 0
Boronic acid-containing starting material
(AACPBA)
t = 40 min t = 160 min Boron-free by-product
Boronic acid-containing product
Absorbance at 254 nm
Fluorescence at 610 nm
Boron-freeby-productBoronic acid-containing
product
AACPBA
Selective on-line detection of boronic acids in HPLC eluates
91
4 Conclusion
An easy-to-use HPLC method was developed for the on-line detection by induced
fluorescence of boronic acids and derivatives using alizarin. The method allows the specific
detection of a wide range of boronic acids and derivatives, even in crude mixtures from
organic synthesis. Boron-free compounds do not give rise to this induced fluorescence.
Through optimization experiments, reasoning and limitations regarding the available
instrumentation, choices were made for the set-up and conditions. Possibilities, however,
are not limited to these choices. Alizarin proved better than ARS for the detection of
MPBA, but ARS may be better in other cases. ACN may be replaced by another solvent or
an aqueous buffer (in which case ARS is recommended), as long as it is compatible with
the eluate. The presence of different acids or bases in the eluates may require a slight
adjustment of the basicity of the alizarin solution (as was done when the eluate contained
0.1% acetic acid). Practical use, which does not always require the highest sensitivity, is not
limited to the optimized parameters such as coil length, coil temperature and alizarin flow
rate (in relation to HPLC flow rate). As such, the present method has shown good potential
for pinpointing boronic acids and derivatives in complex mixtures: it can be of great help,
for example, when one needs to get insight into the course of chemical reactions that
involve boronic acids, or to select the right peak in preparative HPLC separations of
boronic acid-containing mixtures.
Chapter 5
92
5 References
1 C. K. Zacharis and P. D. Tzanavaras, Liquid chromatography coupled to on-line post column
derivatization for the determination of organic compounds: A review on instrumentation and
chemistries, Analytica Chimica Acta, 2013, 798, 1-24.
2 H. A. G. Niederländer, T. A. van Beek, A. Bartasiute and I. I. Koleva, Antioxidant activity assays on-
line with liquid chromatography, Journal of Chromatography A, 2008, 1210, 121-134.
3 A. Dapkevicius, T. A. van Beek and H. A. G. Niederländer, Evaluation and comparison of two
improved techniques for the on-line detection of antioxidants in liquid chromatography eluates, Journal
of Chromatography A, 2001, 912, 73-82.
4 O. Bountagkidou, E. J. C. van der Klift, M. Z. Tsimidou, S. A. Ordoudi and T. A. van Beek, An on-line
high performance liquid chromatography-crocin bleaching assay for detection of antioxidants, Journal
of Chromatography A, 2012, 1237, 80-85.
5 A. Pukalskas, T.A. van Beek, P. de Waard, Development of a triple hyphenated HPLC-radical
scavenging detection-DAD-SPE-NMR system for the rapid identification of antioxidants in complex
plant extracts, Journal of Chromatography A, 2005, 1074, 81-88.
6 M. Pérez-Bonilla, S. Salido, T. A. van Beek, P. J. Linares-Palomino, J. Altarejos, M. Nogueras and A.
Sánchez, Isolation and identification of radical scavengers in olive tree (Olea europaea) wood, Journal
of Chromatography A, 2006, 1112, 311-318.
7 G. Springsteen and B. Wang, A detailed examination of boronic acid-diol complexation, Tetrahedron,
2002, 58, 5291-5300.
8 J. Newman, Novel buffer systems for macromolecular crystallization, Acta Crystallographica Section
D: Biological Crystallography, 2004, 60, 610-612.
93
Chapter 6 General discussion
This chapter provides a reflexion about the work presented in this thesis, suggestions for
future research, and a general conclusion.
Chapter 6
94
1 Immobilization of antibodies using boronic acids: what next?
The oriented immobilization of antibodies using boronic acids, as discussed in chapter 2, is
promising but not yet straightforward. The sugars in the Fc chain of the antibody (N-
glycans) need to expose diols that have enough affinity with boronic acids. The binding
conditions need to be chosen in such a way that the antibody is immobilized on the surface
via the reaction between boronic acid and diol, and not via other types of interactions. The
surface needs to bear long hydrophilic chains to which suitable boronic acids are attached,
for strong and selective binding with the N-glycans. When all these requirements are met,
problems caused by the reversibility of the reaction between boronic acid and diol, thus
release of the antibody from the surface, may still be encountered. In chapter 3, the idea of
making the antibody immobilization irreversible is developed, and it has also been
published by Adak et al.1 A molecule containing an amine, a boronic acid and a diazirine is
attached to the surface via its amine. The antibody binds to the boronic acid via a diol in its
N-glycans, the sample is irradiated with UV light (365 nm) which causes transformation of
the diazirine into a carbene, and this carbene inserts into a C-C, C-H or X-H (X being a
heteroatom) bond in its vicinity, resulting in oriented and irreversible antibody
immobilization.
This novel antibody immobilization strategy looks promising, and it may further be
optimized in future research. In the article of Adak et al., only one trifunctional linker was
tested. However, we expect that the antibody orientation and density on the surface strongly
depend on the structure of the linker. The whole linker (as well as the spacer between linker
and surface) needs to be hydrophilic enough to prevent hydrophobic interactions with the
antibody, and preferably electrically neutral (at the desired pH) to minimize electrostatic
interactions with the antibody (although electrostatic interactions may facilitate antibody
orientation in some cases). In order to choose the most suitable boronic acid structure,
stability of the ester resulting from the binding between boronic acid and diol is no longer
the dominant issue (as opposed to immobilization without diazirine) but most importantly,
the boronic acid needs to bind specifically to a diol in the antibody N-glycans. Secondary
interactions between boronic acid and antibody must be avoided.
General discussion
95
Once the antibody is bound to the surface via the boronic acid, and the sample is irradiated
with UV light, a carbene is formed, which is supposed to react with a part of the antibody in
the vicinity of the N-glycans. The length and flexibility of the spacer between boronic acid
and diazirine play an important role: if the spacer is too short or too stiff, then the carbene
may not be able to access any C-C, C-H or X-H bond of the antibody. If the spacer is
flexible but too long, then the carbene may react with a part of the antibody that is far away
from the N-glycans, resulting in a random orientation, or even with another antibody.
Optimization of the trifunctional linker structure, therefore, may provide even better results
than reported so far.
2 Synthesis of boronic acid-containing compounds
2.1 Boronic acid: to protect or not to protect?
In previous work, before the synthesis of the linkers reported in chapter 3, we protected 4-
carboxyphenylboronic acid (CPBA) with pinacol and 2D-TLC revealed that the obtained
CPBA-pinacol ester degraded on silica. This suggests that synthesized boronate esters
cannot be purified by silica-based chromatography. Moreover, it has been reported that the
boronic acid protecting groups often had to be removed in harsh conditions,2 or were
difficult to separate from the free boronic acids.3 As we had also seen in a literature survey
that unprotected 3-aminophenylboronic acid (APBA) had been used more often than
protected APBA in coupling reactions with carboxylic acids, we decided to work without
protecting groups. Whether this was a good idea is debatable. On the one hand, the use of
free boronic acids saved two steps in the synthetic route (protection and deprotection),
potential problems getting rid of the protecting diol were avoided, and it proved possible to
isolate boronic acid-containing compounds by diol-silica flash chromatography. On the
other hand, a general survey in the Reaxys database (this time, not with APBA only, but
with all boronic acids) showed that boronic esters were used much more often than boronic
acids in coupling reactions between primary amines and carboxylic acids, and in Boc
removal reactions. It is also mentioned in the literature that boronic acids are incompatible
Chapter 6
96
with most synthetic reagents,2 and that their purification and characterization is easier when
they are protected.4 In order to protect boronic acids for a better compatibility with
chemical reagents, and isolate them easily, it may be a good option to use a solid support
that bears diols, as reported by several groups.5-7 Solid-phase synthesis presents the
disadvantage that the reaction has to be clean and complete, but after optimization of the
reaction conditions, this approach can be promising for the synthesis of boronic acid-
containing compounds.
2.2 Boronic acid and thiol-ene reaction: a good combination?
Figure 1. Thiol-ene reaction between Boc-Cys and AACPBA, as performed in chapters 3 and 5.
The reaction between Boc-Cys and AACPBA in the presence of DMPA (Figure 1) was
reported in chapters 3 and 5. In chapter 3, the reaction was first monitored by TLC, which
suggested complete conversion of AACPBA after 45 min. In subsequent experiments, the
reaction was run for 2 h to ensure complete conversion of AACPBA. When the material
was purified once, by diol-silica chromatography (without water), Boc-Cys-AACPBA was
obtained in 88% yield. When the material went through several purification steps by
reversed-phase chromatography (with water), ESI-HRMS suggested that the desired
product had been oxidized: a phenol instead of a boronic acid, and a sulfoxide instead of a
thioether were observed. In chapter 5, the same reaction was monitored by HPLC using a
detection method that differentiates boronic acids from other compounds. The peak
corresponding to AACPBA decreased over time until it disappeared after about 40 min
(which is in accordance with previous TLC monitoring), while a new peak corresponding to
a boronic acid (most probably Boc-Cys-AACPBA) increased. Whereas this observation
General discussion
97
was in line with our expectations, from the moment that AACPBA was completely
consumed, the concentration of Boc-Cys-AACPBA started decreasing again and
meanwhile, a boron-free compound increased dramatically in concentration. Putting all
these observations together, we propose that the following happened: the thiol-ene reaction
took place as expected (in pure and dry THF), and when a sample from the reaction mixture
was diluted 100 times in MeOH/water 1:1 for HPLC injection, Boc-Cys-AACPBA started
being oxidized, to an extent that depended on the concentration of alkene (AACPBA)
present. The reason for this is unclear, but it may have to do with radicals present in the
reaction mixture, which first react with the alkene, and second cause the transformation of
water into hydrogen peroxide, which in turn causes oxidation of the boronic acid into a
phenol4 and oxidation of the thioether into a sulfoxide. Other factors may play a role, as no
thiol-ene reaction in the presence of boronic acids or esters can be found in the Reaxys
database (at the time of writing). However, we showed that this type of reaction is possible,
as we were able to obtain Boc-Cys-AACPBA in a satisfying yield and purity. The yield and
purity may be improved further by stopping the reaction immediately after complete
consumption of the alkene (AACPBA), and avoiding contact with water as much as
possible.
3 Detection of boronic acids using alizarin
3.1 TLC staining reagents for boronic acids: which one is best?
A staining reagent for the detection of boronic acids on silica TLC plates was developed, as
reported in chapter 4. It is based on the reaction between boronic acid and alizarin to form a
fluorescent boronate ester, which can be detected under ambient light (yellow spot on pink
background) or 365 nm light (fluorescent yellow spot on dark background). The staining
reagent simply consists of 1 mM alizarin in acetone, it is selective for boronic acids and
derivatives and detects all of them. This alizarin solution has been used extensively in our
labs, and proved very useful. Less than a month after acceptance of our paper, Lawrence et
al. published another method for the TLC detection of boronic acids.8 TLC plates are
Chapter 6
98
dipped into an acidic curcumin solution in ethanol, and then heated with a heat-gun.
Complexation of yellow curcumin with boronic acids forms red/orange spots. However, we
think that our alizarin-based TLC stain is better than the curcumin-based one: 1) the stain
solution is simpler and safer as it does not require 2 M HCl; 2) the stain solution is
approximately 1000 times cheaper in terms of dye use; 3) it does not require any heating; 4)
the color difference is more obvious (Figure 2); 5) last but not least, it is approximately 10
times more sensitive for the detection of phenylboronic acid (PBA) (Figure 2). The
following year, Aronoff et al. reported a new TLC staining reagent for boronic acids, based
on the binding between 10-hydroxybenzo[h]quinolone and boronic acids.9 An apparent
comparison between the authors’ method and ours is reported, and from these results, the
authors claim their method to be more sensitive than ours. However, our method was
actually not reproduced, as alizarin red S (ARS) was used instead of alizarin. In organic
solvents such as acetone, ARS is poorly soluble and performs poorly, unlike alizarin. By
looking at the real comparison between both methods (Figure 2), it seems that they are
equivalent in terms of sensitivity. However, the alizarin staining reagent presents the
advantages that it is approximately ten times cheaper in terms of dye use and does not
require any heating. We were pleased to see that our method was reported in a book chapter
by Spangenberg on specific TLC staining reagents (in German),10 and used by Marques et
al. for monitoring the conversion of aryl halides into boronic esters.11
Figure 2. Detection of phenylboronic acid using different staining methods. In part from ref. 8 with permission of
The Royal Society of Chemistry and from ref. 9 with permission of the American Chemical Society.
10 1 0.1 0.01PBA conc. (mM)
Alizarin (chapter 4 of this thesis)
Curcumin (Lawrence et al.8)
HBQ (Aronoff et al.9)
ARS (Aronoff et al.9)
Stai
ning
met
hod Exact PBA
conc. (mM)
General discussion
99
3.2 TLC staining reagents for boronic acids: other TLC plates?
After our article was published, we realized that the alizarin staining reagent had not been
tested on other supports than silica plates. In the course of synthetic work, it turned out that
1 mM alizarin in acetone performed very poorly for the detection of boronic acids on
reversed-phase (C18) TLC plates. However, while developing the method for the HPLC on-
line detection of boronic acids using alizarin (as reported in chapter 5), we found that the
addition of 0.1% triethylamine (TEA) to 1 mM alizarin in ACN dramatically improved the
detection of PBA in solution. These results gave the idea of an additional TLC experiment:
solutions of 1 mM alizarin in acetone without and with 0.1% TEA were compared for the
detection of PBA on various TLC plates (Figure 3). The presence of TEA did not have
much effect with silica plates, but it considerably improved the detection of PBA on
reversed-phase (C18 and C8) plates and chromatography paper. Thus, a simple adjustment of
the alizarin solution makes it usable for other TLC plates than silica (except alumina
Al2O3), and fine-tuning of the TEA concentration may further improve the sensitivity of the
method.
Figure 3. Effect of triethylamine on the detection of PBA. Selections from the original photograph of various TLC
plates spotted with PBA (10 mM in MeOH, 2 µL), briefly immersed in 1 mM alizarin solution in acetone (without
or with 0.1% triethylamine), left to dry and placed under 365 nm light.
Silica on glass (Merck)
Silica on aluminum (Merck)
Silica on PET (Fluka)
Al2O3 on aluminum (Merck)
C18-coated silica on
glass
C8-coated silica on
glass
Chromatogr. paper
1 mM alizarinin acetone
1 mM alizarinand 0.1% triethylaminein acetone
Chapter 6
100
3.3 HPLC: why not alizarin red S?
In chapter 5, the development of a method for the HPLC on-line detection of boronic acids
using alizarin was shown. In preliminary off-line experiments, ARS was used because it is
much more commonly used than alizarin. When the detection of various boronic acids was
tested with ARS, very weak signals were observed in the case of electron-donating boronic
acids, especially 4-methoxyphenylboronic acid (MPBA). Therefore, ARS and alizarin were
compared for the detection of MPBA, and the change in fluorescence intensity was much
higher with alizarin than with ARS. From this result, we decided to continue developing the
method with alizarin.
After our work on the HPLC detection of boronic acids, we thought it might be interesting
to compare alizarin and ARS in a more systematic way. Our results had shown that alizarin
was better than ARS for the detection of MPBA, but a literature survey had shown that
ARS had almost always been preferred over alizarin. A comparison of ARS and alizarin
may, therefore, promote the use of alizarin for the detection of boronic acids. Alizarin is
known to be “insoluble” in water, but at a concentration of ~ 10 µM in PBS, solubility did
not cause a problem in our hands. Solutions of alizarin or ARS (10 µM) in PBS pH 7.4,
with or without boronic acid (100 µM), were analyzed by fluorescence spectroscopy.
Various boronic acids were analyzed in this way. After subtracting the background
fluorescence of alizarin and ARS themselves from the excitation and emission spectra of
the mixtures, it was confirmed that alizarin performed better than ARS for the detection of
MPBA. However, the obtained spectra revealed that all the other tested boronic acids
resulted in higher fluorescence intensities with ARS than with alizarin. It turned out that
MPBA was a very particular case on which we had, by chance, based our choice of alizarin
over ARS for the development of the HPLC method.
Further comparative experiments consisted of the detection of 100 µM PBA by 10 µM
alizarin or ARS in various conditions, again by fluorescence spectroscopy. When the pH
was varied from 4 to 10, without any organic solvent, the optimum pH was between 7 and 8
in both cases, but ARS performed slightly better than alizarin in terms of PBA detection. In
PBS without organic solvent, or with 1% MeOH, acetone, THF, ACN, DMF or DMSO,
General discussion
101
ARS still performed slightly better than alizarin. However, in PBS with 50% of the same
solvents, alizarin performed slightly better than ARS. To further investigate the effect of the
percentage of organic solvent, the detection of 100 µM PBA by 10 µM alizarin or ARS was
tested in PBS/MeOH and PBS/ACN mixtures with MeOH or ACN percentages of 10%,
50% and 90%. As a result, as the fraction of organic solvent increased, the fluorescence
intensity increased when alizarin was used and decreased when ARS was used. With 90%
organic solvent (and also 50% in the case of ACN), alizarin proved better than ARS for the
detection of PBA.
To summarize, our comparative study has shown the following: 1) ARS results in higher
changes in fluorescence than alizarin for the detection of boronic acids in aqueous solutions
(possibly with a small percentage of organic solvent), except for the detection of MPBA; 2)
alizarin results in higher changes in fluorescence than ARS for the detection of PBA (and
probably other boronic acids) in solutions with high organic solvent content.
To come back to our work on the on-line HPLC detection of boronic acids as described in
chapter 5, the question is: what is better, alizarin or ARS? We used a solution of 75 µM
alizarin and 0.1% triethylamine in ACN. Because we made the choice of alizarin over ARS
in preliminary off-line experiments, ARS was not tested on-line. Considering the
experiments described above, a buffered aqueous solution of ARS at pH 7-8 may result in a
higher sensitivity for the on-line detection of boronic acids. However, mixing the HPLC
eluate with the ARS solution may cause problems: salts from the buffered aqueous solution
of ARS may precipitate when mixing with an organic solvent from the eluate, organic
molecules in the eluate may precipitate when mixing with the aqueous buffer, and bubbles
may be formed upon mixing of MeOH from the eluate and water from the ARS solution.
In conclusion of this section, both alizarin and ARS may be considered for the HPLC on-
line detection of boronic acids, depending on the experimental conditions. If the HPLC
eluate contains no or little organic solvent, then ARS in an aqueous buffer may be
preferred. If the HPLC eluate contains a high proportion of organic solvent, alizarin in an
organic solvent is the way to go.
Chapter 6
102
4 Separation of boronic acids by diol-silica chromatography
The isolation of boronic acid-containing compounds by diol-silica flash chromatography
was reported for the first time (to the best of our knowledge) in chapter 3. Commercially
available diol-silica (structure page 43) was used in the same way as bare silica in a
traditional flash chromatography setup. The eluent was a gradient of MeOH (and acetic
acid in one case) in EtOAc which enabled, first, the elution of boronic acid-free compounds
and second, the elution of boronic acid-containing compounds. Only two compounds were
purified in this way, and this separation method deserves further investigation.
Separation modes. Diol-silica has previously been used, as discussed in chapter 3, for
separations based on polarity (normal-phase and reversed-phase chromatography) or
hydrophilicity (hydrophilic interaction chromatography) of the analytes. In addition, diol-
silica may be used for separations based on the reversible reaction between diols of the
stationary phase and boronic acid groups of the analytes. The separation of boronic acids
may be based on one or several of these separation modes (polarity, hydrophilicity and
boronic acid-diol interaction), depending on the structure of these compounds and the
composition of the mobile phase. Thus, it would be interesting to investigate the influence
of these factors on the separation of boronic acids.
Scope. Until now, only two boronic acids were isolated using diol-silica. The results look
promising, but they are just a starting point as they cannot tell whether this purification
method is generally applicable. Further investigations, therefore, may include the
purification of a wide range of boronic acids.
Diol-silica HPLC. The separation of boronic acids by a diol-silica stationary phase may
further be investigated by HPLC, using a commercially available diol-silica HPLC column.
HPLC would enable rapid screening of boronic acids and mobile phases, provide
information more clearly and more efficiently than “manual” column chromatography
about scope and separation modes, and expand the applicability of this unexplored
separation strategy. Moreover, HPLC analysis would be a first step towards the separation
General discussion
103
of boronic acids by preparative diol-silica HPLC, which may be a solution to the difficult
purification of unprotected boronic acids.
5 General conclusion
The initial goal of the PhD project was to functionalize sensor chips in such a way that
biomarkers could be detected at low concentrations (~ 1 pM) in urine. This would help to
enable an objective and reliable diagnosis of depression. Antibodies would be immobilized
on the chips in an oriented way, with their antigen binding sites pointing away from the
surface. With this goal in mind, the immobilization of antibodies via boronic acids was
explored. A strategy for the oriented and irreversible immobilization of antibodies was
devised, and relevant boronic acid-containing linkers were synthesized. This strategy and
its proof-of-principle were published by another group at the moment our linkers were
obtained. Eventually, the initial goal of the PhD project was not achieved within the
available time, for several reasons that were out of our hands. However, the challenges
encountered during our work with boronic acids gave rise to valuable findings. Some
purification experiments suggested that boronic acids can be isolated by means of diol-
silica chromatography (chapter 3). Boronic acids can be detected on TLC plates by dipping
these plates in an alizarin solution and visualizing them under a UV lamp (365 nm), as
shown in chapter 4. This is the first published TLC staining method for boronic acids. It is
selective, sensitive, simple, safe and inexpensive, and by these combined advantages, we
find it better than alternative, even more recently published, TLC staining methods for
boronic acids. Moreover, as shown in chapter 5, boronic acids can also be detected on-line
after HPLC separation, again by using alizarin. The HPLC eluate mixes with an alizarin
solution, and fluorescent complexes are produced when boronic acids are present. This
facilitates the pinpointing of boronic acids in complex mixtures, and enables the
optimization of conditions for the isolation of boronic acids by preparative HPLC. As such,
the synthetic potential of boronic acids may be further developed, as detection of these
compounds and derivatives thereof is significantly facilitated by the work presented in this
thesis.
Chapter 6
104
6 References
1 A. K. Adak, B.-Y. Li, L.-D. Huang, T.-W. Lin, T.-C. Chang, K. C. Hwang and C.-C. Lin, Fabrication
of antibody microarrays by light-induced covalent and oriented immobilization, ACS Applied
Materials & Interfaces, 2014, 6, 10452-10460.
2 E. P. Gillis and M. D. Burke, Multistep synthesis of complex boronic acids from simple MIDA
boronates, Journal of the American Chemical Society, 2008, 130, 14084-14085.
3 C. Malan, C. Morin and G. Preckher, Two reducible protecting groups for boronic acids, Tetrahedron
Letters, 1996, 37, 6705-6708.
4 D. G. Hall, Structure, properties, and preparation of boronic acid derivatives. Overview of their
reactions and applications, John Wiley & Sons: Weinheim, Germany, 2006.
5 D. G. Hall, J. Tailor and M. Gravel, N,N-diethanolaminomethyl polystyrene: An efficient solid support
to immobilize boronic acids, 1999, 38, 3064-3067.
6 W. Yang, X. Gao, G. Springsteen and B. Wang, Catechol pendant polystyrene for solid-phase
synthesis, Tetrahedron Letters, 2002, 43, 6339-6342.
7 B. Carboni, C. Pourbaix, F. Carreaux, H. Deleuze and B. Maillard, Boronic ester as a linker system for
solid phase synthesis, Tetrahedron Letters, 1999, 40, 7979-7983.
8 K. Lawrence, S. E. Flower, G. Kociok-Kohn, C. G. Frost and T. D. James, A simple and effective
colorimetric technique for the detection of boronic acids and their derivatives, Analytical Methods,
2012, 4, 2215-2217.
9 M. R. Aronoff, B. VanVeller and R. T. Raines, Detection of boronic acids through excited-state
intramolecular proton-transfer fluorescence, Organic Letters, 2013, 15, 5382-5385.
10 B. Spangenberg, Spezifische Färbereagenzien, in Quantitative Dünnschichtchromatographie, Springer,
Berlin Heidelberg, 2014, 171-226.
11 C. S. Marques, D. Peixoto and A. J. Burke, Transition-metal-catalyzed intramolecular cyclization of
amido(hetero)arylboronic acid aldehydes to isoquinolinones and derivatives, RSC Advances, 2015, 5,
20108-20114.
105
Appendix Information supplementing chapter 5
Appendix
106
Position Material o.d. (mm)
i.d. (mm)
Length (cm) Supplier Part number
From solvent reservoirs to HPLC delivery system PTFE 2 90 Gilson 3645357
From HPLC delivery system to injector PEEK 0.13 20 Vici JR-T-6007
From injector to column - -
From column to absorbance detector
Stainless steel 1.59 0.25 30 Grace 2106936
From absorbance detector to T-piece PEEK 1.59 0.25 15 Vici JR-T-6009
From Superloop to T-piece PTFE 1.59 0.76 15 Grace 35670
From HPLC pump to Superloop PTFE 1.59 0.76 85 Grace 35670
From T-piece to fluorescence detector PEEK 0.25 Var.* Vici JR-T-6009
From fluorescence detector to waste PTFE 1.59 0.76 100 Grace 35670
Table 1. Tubing used in the on-line HPLC instrumental set-up.*Several reaction coil lengths were tested, and a
length of 350 cm was selected.
Information supplementing chapter 5
107
Figure 1. Analysis of 3.13 μM ARS and 3.13 μM ARS + 313 μM PBA solutions in PBS pH 7.4. A. Excitation
spectra, λem 610 nm, and subtracted spectrum (difference between fluorescence intensities of ARS and ARS-PBA).
B. Emission spectra, λex 469 nm, and subtracted spectrum.
0E+0
2E+4
4E+4
6E+4
8E+4
350 400 450 500 550
Fluo
resc
ence
Inte
nsity
(a.u
.)
Excitation wavelength (nm)
ARS
ARS-PBA
ARS-PBA minus ARS
A
0E+0
1E+5
2E+5
3E+5
480 530 580 630 680
Fluo
resc
ence
Inte
nsity
(a.u
.)
Emission wavelength (nm)
B
Appendix
108
Figure 2. Analysis of 3.13 μM ARS and 3.13 μM ARS + 313 μM solutions of various boronic acids in PBS pH
7.4. A. Subtracted (see explanation in Figure 1) excitation spectra, λem 610 nm. B. Subtracted emission spectra, λex
469 nm.
0E+0
2E+4
4E+4
6E+4
8E+4
1E+5
350 400 450 500 550
Fluo
resc
ence
Inte
nsity
(a.u
.)
Excitation wavelength (nm)
4-Trifluoromethylphenylboronic acid4-Carboxyphenylboronic acidPhenylboronic acidButylboronic acid3-Aminophenylboronic acid4-Methoxyphenylboronic acid
A
0E+0
1E+3
2E+3
3E+3
350 400 450 500 550
Fluo
resc
ence
Inte
nsity
(a.u
.)
Excitation wavelength (nm)
Information supplementing chapter 5
109
Figure 2. (Continued).
0E+0
1E+5
2E+5
3E+5
4E+5
5E+5
480 530 580 630 680
Fluo
resc
ence
Inte
nsity
(a.u
.)
Emission wavelength (nm)
4-Trifluoromethylphenylboronic acid4-Carboxyphenylboronic acidPhenylboronic acidButylboronic acid3-Aminophenylboronic acid4-Methoxyphenylboronic acid
B
0E+0
1E+4
480 530 580 630 680
Fluo
resc
ence
Inte
nsity
(a.u
.)
Emission wavelength (nm)
Appendix
110
Figure 3. Comparison of 3.13 μM ARS and 3.13 μM alizarin for the detection of 313 μM PBA in ACN/water 1:1.
A. Excitation spectra, λem 610 nm, and subtracted spectrum (difference between fluorescence intensities of ARS
and ARS-MPBA or ALZ and ALZ-MPBA) . B. Emission spectra, λex 469 nm, and subtracted spectra.
0E+0
1E+4
2E+4
3E+4
350 400 450 500 550
Fluo
resc
ence
inte
nsity
(a.u
.)
Excitation wavelength (nm)
ARSARS-MPBAARS-MPBA minus ARSALZALZ-MPBAALZ-MPBA minus ALZ
A
0E+0
1E+5
2E+5
480 530 580 630 680
Fluo
resc
ence
inte
nsity
(a.u
.)
Emission wavelength (nm)
B
Information supplementing chapter 5
111
Figure 4. Effect of solvent composition on the detection of 100 μM PBA using 10 μM alizarin. Various
percentages of ACN (1) or MeOH (2) were used. The remaining volumes consisted of phosphate buffered saline
pH 7.4. A. Subtracted (see explanation in Figure 1) excitation spectra, λem 610 nm. B. Subtracted emission spectra,
λex 469 nm.
0E+0
1E+4
2E+4
3E+4
4E+4
5E+4
350 400 450 500 550
Fluo
resc
ence
Inte
nsity
(a.u
.)
Excitation wavelength (nm)
90% ACN
50% ACN
10% ACN
A1
0E+0
1E+4
2E+4
3E+4
4E+4
5E+4
350 400 450 500 550
Fluo
resc
ence
Inte
nsity
(a.u
.)
Excitation wavelength (nm)
90% MeOH
50% MeOH
10% MeOH
A2
Appendix
112
Figure 4. (Continued).
0E+0
1E+5
480 530 580 630 680
Fluo
resc
ence
Inte
nsity
(a.u
.)
Emission wavelength (nm)
90% ACN
50% ACN
10% ACN
B1
0E+0
1E+5
480 530 580 630 680
Fluo
resc
ence
Inte
nsity
(a.u
.)
Emission wavelength (nm)
90% MeOH
50% MeOH
10% MeOH
B2
Information supplementing chapter 5
113
Figure 5. Effect of pH on the detection of 100 μM PBA using 10 μM alizarin, in 20 mM citrate/30 mM HEPES/40
mM CHES buffer at various pH values and identical organic compositions. λex 469 nm, λem 610 nm.
0,0E+0
5,0E+3
1,0E+4
1,5E+4
4 5 6 7 8 9 10
Fluo
resc
ence
inte
nsity
(a.u
.)
pH
ALZ
ALZ-PBA
ALZ-PBA minus ALZ
Appendix
114
Figure 6. Effect of TEA on the detection of 100 μM PBA using 10 μM alizarin. Solutions consisted of two parts of
equal volumes: 1) “alizarin solution mimic”, 20 μM alizarin solution in ACN that contained 0%, 0.1%, 1% or 10%
TEA, 2) “HPLC eluate mimic”, 200 μM PBA in 50% water and 50% of ACN (A) or MeOH (B). λex 469 nm, λem
610 nm.
0E+0
2E+4
4E+4
6E+4
8E+4
0 0.1 1 10
Fluo
resc
ence
inte
nsity
(a.u
.)
TEA % in alizarin solution
ALZ
ALZ-PBA
ALZ-PBA minus ALZ
A
0E+0
2E+4
4E+4
6E+4
8E+4
0 0.1 1 10
Fluo
resc
ence
inte
nsity
(a.u
.)
TEA % in alizarin solution
B
Information supplementing chapter 5
115
Figure 7. Effect of TEA on the detection of 100 μM PBA using 10 μM alizarin, in presence of acetic acid.
Solutions consisted of two parts of equal volumes: 1) “alizarin solution mimic”, 20 μM alizarin solution in ACN
that contained 0%, 0.1%, 1% or 10% TEA, 2) “HPLC eluate mimic”, 200 μM PBA and 0.1% acetic acid in 50%
water and 50% of ACN (A) or MeOH (B). λex 469 nm, λem 610 nm.
0E+0
2E+4
4E+4
6E+4
8E+4
0 0.1 1 10
Fluo
resc
ence
inte
nsity
(a.u
.)
TEA % in alizarin solution
ALZ
ALZ-PBA
ALZ-PBA minus ALZ
A
0E+0
2E+4
4E+4
6E+4
8E+4
0 0.1 1 10
Fluo
resc
ence
inte
nsity
(a.u
.)
TEA % in alizarin solution
B
Appendix
116
Figure 8. Effect of the coil temperature on the height of the fluorescent peak produced by the reaction between
alizarin and the injected PBA.
Figure 9. Effect of the alizarin concentration on the signal-to-noise ratio of the peaks produced by the reaction
between alizarin and the three lowest concentrations of injected PBA. n = 3 except for 50 µM alizarin/all PBA
concentrations and 75 µM alizarin/3 µM PBA in which case n = 2.
y = 0.1906x + 1.0576
R² = 0.9995
0
5
10
20 30 40 50
Peak
hei
ght (
mA
U)
Temperature (°C)
0
10
20
30
40
50
60
70
1 3 10 30 50 75 100 300
S/N
Alizarin concentration (μM)
1
3
10
PBA conc.(μM)
Information supplementing chapter 5
117
Figure 10. Chromatograms for the detection of various concentrations of PBA using the developed method.
Results from paragraph 3.3.1 in chapter 5, for an alizarin concentration of 75 µM.
20
30
40
50
3.0 3.5 4.0 4.5 5.0
mA
U
min
20
30
40
50
3.0 3.5 4.0 4.5 5.0
mA
U
min
20
30
40
50
3.0 3.5 4.0 4.5 5.0
mA
U
min
300 µM PBA
100 µM PBA
30 µM PBA
Appendix
118
Figure 10. (Continued).
22.8
23.8
3.0 3.5 4.0 4.5 5.0
mA
U
min
22.8
23.8
3.0 3.5 4.0 4.5 5.0
mA
U
min
22.8
23.8
3.0 3.5 4.0 4.5 5.0
mA
U
min
10 µM PBA (zoom)
3 µM PBA (zoom)
1 µM PBA (zoom)
Information supplementing chapter 5
119
Figure 11. Ability of the developed method to quantify PBA. Peak areas taken from chromatograms in Figure S10
(+ replicates). n = 3 except for 3 µM PBA in which case n = 2. A. Peak area for all PBA concentrations. B. Peak
areas for lower PBA concentrations (1, 3 and 10 μM).
y = 1.5326xR² = 0.9997
0
100
200
300
400
500
0 100 200 300
Peak
are
a
PBA concentration (µM)
A
y = 1.1073xR² = 0.9925
0
2
4
6
8
10
12
0 5 10
Peak
are
a
PBA concentration (µM)
B
Appendix
120
Figure 12. Detection of various boronic acids using the developed method.
0
20
40
5 6 7 8 9
mA
U
min
0
20
40
5 6 7 8 9
mA
U
min
0
20
40
5 6 7 8 9
mA
U
min0
20
40
5 6 7 8 9
mA
U
min
0
20
40
5 6 7 8 9
mA
U
min
0
20
40
5 6 7 8 9m
AU
min
Phenylboronic acid, 300 μM Phenylboronic acid pinacol ester, 300 μM
Phenylboronic acid MIDA ester, 300 μM Potassium phenyltrifluoroborate, 300 μM
Butylboronic acid, 1320 μM Potassium butyltrifluoroborate, 1200 μM
BHO OH
BOO
BOHHO
B FF
F
K+
Fluorescence detector with alizarin Fluorescence detector without alizarin Absorbance detector
Information supplementing chapter 5
121
Figure 12. (Continued).
0
20
40
60
2 3 4 5 6
mA
U
min
0
20
40
60
12 13 14 15 16
mA
U
min0
20
40
60
5 6 7 8 9
mA
U
min
0
20
40
60
0 1 2 3 4
mA
U
min
4-Trifluoromethyl-phenylboronic acid, 250 μM
4-Methoxyphenylboronic acid, 300 μM
3-Aminophenylboronic acid, 2500 μM
BHO OH
COOH
BHO OH
CF3
BHO OH
OCH3
BOHHO
NH2
4-Carboxyphenylboronic acid, 300 μM
Fluorescence detector with alizarin Fluorescence detector without alizarin Absorbance detector
Appendix
122
Figure 13. Detection of the autofluorescent boron-free compounds using the developed method.
Figure 14. Monitoring of the acid-catalyzed esterification of CPBA. At each reaction time, fluorescence peak
heights that correspond to starting material CPBA (4.6 min) and product (22.6 min) were determined.
0
50
100
150
200
0 0.5 1 2 3 6 24
Peak
hei
ght (
mA
U)
Reaction time (hours)
Starting material (CPBA)
Product
0
20
40
60
80
2 3
mA
U
min0
20
40
60
80
2 3
mA
U
min0
20
40
60
80
2 3
mA
U
min0
20
40
60
80
2 3
mA
U
min0
20
40
60
80
2 3
mA
U
min
Fluorescence detector with alizarin Fluorescence detector without alizarin Absorbance detector
OH
NH2O
3.00 mM 3.00 mM 0.30 mM 0.30 mM 0.30 mM
Fluorescence detector without alizarin
Information supplementing chapter 5
123
Figure 15. Monitoring of the thiol-ene reaction between AACPBA and Boc-Cys. At each reaction time, peak
heights that correspond to starting material (3.2 min by absorbance), by-product (7.0 min by absorbance) and
product (8.1 min by absorbance) were determined. A: absorbance detection at 254 nm. B: fluorescence detection at
610 nm.
0
50
100
150
200
250
0 1 2 5 10 20 40 80 160
Peak
hei
ght (
mA
U)
Reaction time (min)
Starting material (AACPBA)
By-product
Product
A
0
50
100
150
200
250
0 1 2 5 10 20 40 80 160
Peak
hei
ght (
mA
U)
Reaction time (min)
B
124
125
Summary
Chapter 1 introduces the theory and known applications of the interaction between boronic
acids and diols, and explains the context of this thesis. Diagnosis of depression was the
initial goal of this multidisciplinary project. The focus of the PhD project was the
development of a strategy to immobilize antibodies on the surface of a chip in such a way
that very low concentrations (~ 1 pM) of biomarkers for depression could be detected in
urine. To achieve this, the immobilization of antibodies using boronic acids seemed
promising.
However, preliminary experiments and further insights revealed the many challenges that
this immobilization strategy faces, giving rise to Chapter 2. This chapter discusses several
important points that need to be taken into account when one plans to immobilize
antibodies via boronic acids: choice of the boronic acid structure and spacer to attach it to
the surface, use of an antifouling polymer, choice of an antibody with suitable
glycosylation, optimization of the conditions for antibody immobilization...
One big issue for antibody immobilization using boronic acids is the reversibility of the
reaction between boronic acids and diols, hence the possible release of the antibody from
the surface. Chapter 3 describes the design and synthesis of boronic acid-containing
linkers that would enable the oriented and irreversible immobilization of antibodies. Two
linkers were designed with an amine for surface attachment, a boronic acid for capturing
antibodies via the N-glycans in their Fc chain, and a diazirine for irreversible
immobilization upon UV irradiation while maintaining antibody orientation. From a
diazirine building-block that was obtained in three steps, the first linker was synthesized in
four steps and the second linker was synthesized in three steps. Diol-functionalized silica
was used for the chromatography of two boronic acid-containing intermediates, this method
being novel (to the best of our knowledge) and likely based on boronic acid-diol
interactions. High-resolution mass spectrometry, through matching exact masses, matching
isotope patterns and observation of species corresponding to the esterification of boronic
acids with MeOH, confirmed that both linkers were synthesized successfully.
126
During the synthesis of boronic acid-containing linkers, it was difficult to see which spots
on TLC plates corresponded to boronic acids. To solve this problem, a new TLC staining
method based on the reaction between boronic acids and alizarin was developed. Chapter 4
presents this work in detail. After optimization experiments, 1 mM alizarin in acetone was
shown to be the preferred staining solution. When the TLC plate was briefly dipped in this
solution, allowed to dry in ambient air and observed under 365 nm light, bright yellow
fluorescent spots were observed where boronic acids were present. Phenylboronic acid was
detected at a concentration as low as 0.1 mM. A range of boronic acids and derivatives was
successfully detected, and boron-free compounds resulted in no or very weak fluorescence.
The staining method was further tested in the monitoring of three reactions involving
boronic acids, and provided clear information about the consumption or formation of
boronic acid-containing compounds.
Although TLC is useful to synthetic chemists, analysis of reaction mixtures by HPLC is
sometimes necessary for obtaining more accurate information or for optimization of
preparative HPLC conditions. Chapter 5 presents the development and applicability of a
method for the on-line HPLC detection of boronic acids using alizarin. After optimization
experiments at an HPLC flow rate of 0.40 mL/min, the HPLC-separated analytes were
mixed post-column with a solution of 75 μM alizarin and 0.1% triethylamine in ACN,
which was delivered at a flow rate of 0.60 mL/min. The reaction between alizarin and
boronic acids occurred in a reaction coil of dimensions of 3.5 m × 0.25 mm at a
temperature of 50 °C, resulting in fluorescent complexes that were detected as positive
peaks by a fluorescence detector (exc 469 nm and em 610 nm). The method enabled the
selective detection of various boronic acids and derivatives, with a limit of detection of
phenylboronic acid of 1.2 ng or 1 μM. It could successfully monitor the progress of two
organic reactions involving boronic acid-containing compounds, and provided useful
insights into the course of the reactions.
Chapter 6 provides a reflexion about the work presented in this thesis, suggestions for
future research, and a general conclusion.
127
Acknowledgements
Teris van Beek, Han Zuilhof, Alexandre Villela, Radostina Manova, Elly Geurtsen,
Elbert van der Klift, Cees van de Haar, Denis Dorokhin, Barend van Lagen, Hendra
Willemen, Frank Claassen, Sidharam (Sidhu) Pujari, Ai Nguyen, Nagendra Bhairamadgi,
Marcel Giesbers, Jacob Baggerman, Anke Trilling, Satesh Gangarapu, Ronald de Bruin,
Willem van Heugten, Sourav Bhattacharjee, Nagesh Wagdare, Ton Marcelis, Aleida
Ruisch, Umesh Chinnaswamy Panchatcharam, Loes Ruizendaal, Jaime Garcia Hartjes, Jos Paulusse, Carel Weijers, Saurabh Srivastava, Tin Doan, Maurice Franssen, Wouter Biesta, Erik van Rozendaal, Mabel Caipa Campos, Anne-Marie Franssen, Kim van
Gessel, Maarten Posthumus, Michel Nielen, Cees van Rijn, Bart Rijksen, Dani Lucas-
Barbosa, Tom Wennekes, Willem Haasnoot, Suzanne Jansze, Aart van Amerongen,
Truus Posthuma-Trumpie, Peter Steunenberg, Esther van Andel, Quirijn van
Dierendonck, Luc Scheres, Yao Shen, Aline Debrassi, Michiel Uiterweerd, Bas van den
Berg, Esther Roeven, Aljosha Filippov, Geert Noordzij, Wilco Duvivier, Sweccha Joshi,
Martine Tiddens, Jorin Hoogenboom, Jose Maria (Txema) Alonso Carnicero, Rickdeb
Sen, Tjerk Sminia, Christie Nguyen, Zhanhua Wang, Carole Messerli, Hanna de Jong,
Jelmer van der Rijst, Iris van Marwijk, Bram Bielen, Maarten Smulders, Sjoerd
Slagman, Puspita Wardani, Farida Kurniawati, Medea Kosian, Pepijn Geutjes, Adriaan
van der Meer, Sven Krabbenborg, Digvijay Gahtory, Laura Schijven, Judith Firet, Fred
van Geenen, Fátima García Melo, Jorge Escorihuela Fuentes, Linda Kaster and Stefanie Lange (in approximate order of appearance in the story behind this thesis), you have all
made contributions to my PhD project, ranging from answering simple questions to getting
into the core of my work. Moreover, I really enjoyed getting to know you and sharing nice
moments with you. Each one of you, thank you!
In particular, I would like to highlight some important contributions to my PhD project,
inside and outside the work described in this thesis. Han, you believed in me from the start
to the end. You encouraged me with your positive attitude, supported and enriched my
ideas. Thank you! Teris, we sat together for many hours, having in-depth discussions on
our work, and you helped me with patience and eye for detail. Thank you! Aart, Truus and
128
Laura, thank you for your commitment to our antibody-related projects. For reasons that
were out of our hands, the obtained results were not satisfying, but many of these results
provided insights that made chapter 2 possible. Txema, Jorge, Laura, Tjerk and Bas,
thank you for revising my manuscript for chapter 2. Suzanne, Martine, Hanna and Iris,
thank you for your hard work on the synthesis of boronic acid-containing linkers for
antibody immobilization. Together we learned how to handle these difficult boronic acids,
giving rise to chapters 3 and 4. Tom, thank you for your help and suggestions with
synthetic issues and your original ideas. Puspita, thank you for your devotion to our HPLC
project, almost one year long. Your family was at the other side of the world, you went
through tough times, but your hard work made chapter 5 possible. Alexandre, thank you
for all our meaningful conversations. From my very first day at ORC, I have learned from
you in many aspects, such as theoretical and practical analytical chemistry, time
management, and especially looking at the positive side of things! Radostina, thank you
for answering my countless questions, especially when I had just started my PhD. Jaime,
thank you for the extensive answers to my questions related to finishing my PhD. Elly and
Aleida, thank you for your help with the administrative tasks. Elbert, thank you for your
multidisciplinary technical help, from urgent product order to special tubing and from
software installation to broken freezer. Barend and Pepijn, thank you for your help with
NMR, UV-vis and fluorescence spectroscopy. Frank and Wilco, thank you for your help
with mass spectrometric issues. Luc, thank you for your practical help with surface
modifications. Sidhu, thank you for analyzing many of my samples by XPS. Bas, you
helped me with thiol-ene chemistry, computer issues and microparticles, just to name a few
things. Also, you accepted to be my paranymph although my defense is just a few weeks
before the end of your contract at ORC. Thank you! Ai, you helped me with questions
about antibodies, other proteins, silane versus alkene chemistry, slide holders and other
things, and you accepted to be my paranymph, although you are a busy mom... thank you!
My Dutch and French families, thank you for your support, and especially Martijn, my
true love, thank you for your patience!
129
About the author’s chemical pathway
Florine Duval was born on 10th December 1985 in Toulouse, France. She grew up happily
in a quiet village with her parents and brother. When she was 10, she knew she wanted to
become a scientist. At the age of 15, she knew she wanted to become a chemist and took a
scientific direction at school. Two years later, organic chemistry already was her subject of
choice. After obtaining a scientific high-school diploma with honors in 2003, she went for
hands-on experience and followed a two-year practical training in chemistry in Université
Paul Sabatier (Castres, France). This training included an internship in Université de
Montréal (Montréal, Canada), where she took her first steps in organic synthesis outside
practical courses. She obtained her “Diplôme Universitaire de Technologie” (DUT) in
2005. This would have enabled her to work as a technician, but she wanted to learn more
theory, entered the “Ecole Nationale Supérieure de Chimie de Montpellier” (ENSCM,
Montpellier, France) and followed a three-year intensive training for future executives in
chemistry. Between the second and third years, she carried out an internship at Pfizer
(Sandwich, UK) for one year, where she synthesized medicinal targets and investigated
regioselective O-alkylations of indazolinone. The third year of training (fifth academic
year) included a six-month project at Mercachem (Nijmegen, The Netherlands), where she
synthesized molecules for a client in pharmaceutical research. After obtaining her
“Diplôme d’Ingénieur Chimiste” and MSc in biomolecular chemistry in 2009, she
continued working at Mercachem for one year, as a “Senior Research Chemist”. In 2010,
she started her PhD under the supervision of Teris van Beek and Han Zuilhof, in the
Laboratory of Organic Chemistry, Wageningen University. The successful results are
presented in this thesis.
130
Publications
A. Randall and F. Duval, Regioselective O-alkylations of indazolinone using
(cyanomethylene) triphenylphosphorane, Synlett, 2009, 16, 2673-2675.
F. Duval, T.A. van Beek and H. Zuilhof, Sensitive thin-layer chromatography detection of
boronic acids using alizarin, Synlett, 2012, 23, 1751-1754.
F. Duval, T.A. van Beek and H. Zuilhof, Key steps towards the oriented immobilization of
antibodies using boronic acids, Analyst, 2015, DOI: 10.1039/C5AN00589B.
F. Duval, P.A. Wardani, H. Zuilhof and T.A. van Beek, Selective on-line detection of
boronic acids and derivatives in high-performance liquid chromatography eluates by post-
column reaction with alizarin, Journal of Chromatography A, revision submitted.
131
Overview of completed training activities
Name of the activity Graduate school/institute Year
Discipline-specific activities
Advanced Organic Chemistry 1&2 VLAG 2010-2014
NWO analytical chemistry conference, Lunteren (2×) (poster) 2010-2012
NWO organic chemistry conference, Lunteren (3×) (poster) 2010-2013
Lab-on-a-Chip European Congress, Hamburg, Germany (poster) 2011
NWO CHAINS conference, Maarssen 2011
NanoSensEU symposium on Biosensor Development, Hasselt, Belgium 2012
KNCV organic chemistry conference, Wageningen 2013
General courses
VLAG PhD week VLAG 2011
Advanced course guide to scientific artwork WUR Library 2011
Techniques for writing and presenting a scientific paper Language Centre 2012
Career perspectives WGS 2013
ACS on Campus event, Utrecht 2013
Effective behaviour in your professional surroundings WGS 2014
Mobilizing your network in 2.5 hours Young AFSG 2014
Optionals
Preparation of PhD proposal 2010
Colloquia ORC 2010-2014
Group meetings ORC 2010-2014
PhD trips: UK in 2011, Germany and Switzerland in 2013 ORC 2011-2013
The research presented in this thesis was financially supported by Pieken in de Delta Oost-Nederland.
Financial support from Wageningen University for printing this thesis is gratefully
acknowledged.
Design and layout by Florine Duval
Printed by GVO drukkers & vormgevers B.V. | Ponsen & Looijen, Ede, The Netherlands